ben.guo
5 months ago
parent
df252a9622
commit
305130130b
@ -0,0 +1,2 @@
|
||||
|
||||
.DS_Store
|
||||
@ -0,0 +1,661 @@
|
||||
GNU AFFERO GENERAL PUBLIC LICENSE
|
||||
Version 3, 19 November 2007
|
||||
|
||||
Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
|
||||
Everyone is permitted to copy and distribute verbatim copies
|
||||
of this license document, but changing it is not allowed.
|
||||
|
||||
Preamble
|
||||
|
||||
The GNU Affero General Public License is a free, copyleft license for
|
||||
software and other kinds of works, specifically designed to ensure
|
||||
cooperation with the community in the case of network server software.
|
||||
|
||||
The licenses for most software and other practical works are designed
|
||||
to take away your freedom to share and change the works. By contrast,
|
||||
our General Public Licenses are intended to guarantee your freedom to
|
||||
share and change all versions of a program--to make sure it remains free
|
||||
software for all its users.
|
||||
|
||||
When we speak of free software, we are referring to freedom, not
|
||||
price. Our General Public Licenses are designed to make sure that you
|
||||
have the freedom to distribute copies of free software (and charge for
|
||||
them if you wish), that you receive source code or can get it if you
|
||||
want it, that you can change the software or use pieces of it in new
|
||||
free programs, and that you know you can do these things.
|
||||
|
||||
Developers that use our General Public Licenses protect your rights
|
||||
with two steps: (1) assert copyright on the software, and (2) offer
|
||||
you this License which gives you legal permission to copy, distribute
|
||||
and/or modify the software.
|
||||
|
||||
A secondary benefit of defending all users' freedom is that
|
||||
improvements made in alternate versions of the program, if they
|
||||
receive widespread use, become available for other developers to
|
||||
incorporate. Many developers of free software are heartened and
|
||||
encouraged by the resulting cooperation. However, in the case of
|
||||
software used on network servers, this result may fail to come about.
|
||||
The GNU General Public License permits making a modified version and
|
||||
letting the public access it on a server without ever releasing its
|
||||
source code to the public.
|
||||
|
||||
The GNU Affero General Public License is designed specifically to
|
||||
ensure that, in such cases, the modified source code becomes available
|
||||
to the community. It requires the operator of a network server to
|
||||
provide the source code of the modified version running there to the
|
||||
users of that server. Therefore, public use of a modified version, on
|
||||
a publicly accessible server, gives the public access to the source
|
||||
code of the modified version.
|
||||
|
||||
An older license, called the Affero General Public License and
|
||||
published by Affero, was designed to accomplish similar goals. This is
|
||||
a different license, not a version of the Affero GPL, but Affero has
|
||||
released a new version of the Affero GPL which permits relicensing under
|
||||
this license.
|
||||
|
||||
The precise terms and conditions for copying, distribution and
|
||||
modification follow.
|
||||
|
||||
TERMS AND CONDITIONS
|
||||
|
||||
0. Definitions.
|
||||
|
||||
"This License" refers to version 3 of the GNU Affero General Public License.
|
||||
|
||||
"Copyright" also means copyright-like laws that apply to other kinds of
|
||||
works, such as semiconductor masks.
|
||||
|
||||
"The Program" refers to any copyrightable work licensed under this
|
||||
License. Each licensee is addressed as "you". "Licensees" and
|
||||
"recipients" may be individuals or organizations.
|
||||
|
||||
To "modify" a work means to copy from or adapt all or part of the work
|
||||
in a fashion requiring copyright permission, other than the making of an
|
||||
exact copy. The resulting work is called a "modified version" of the
|
||||
earlier work or a work "based on" the earlier work.
|
||||
|
||||
A "covered work" means either the unmodified Program or a work based
|
||||
on the Program.
|
||||
|
||||
To "propagate" a work means to do anything with it that, without
|
||||
permission, would make you directly or secondarily liable for
|
||||
infringement under applicable copyright law, except executing it on a
|
||||
computer or modifying a private copy. Propagation includes copying,
|
||||
distribution (with or without modification), making available to the
|
||||
public, and in some countries other activities as well.
|
||||
|
||||
To "convey" a work means any kind of propagation that enables other
|
||||
parties to make or receive copies. Mere interaction with a user through
|
||||
a computer network, with no transfer of a copy, is not conveying.
|
||||
|
||||
An interactive user interface displays "Appropriate Legal Notices"
|
||||
to the extent that it includes a convenient and prominently visible
|
||||
feature that (1) displays an appropriate copyright notice, and (2)
|
||||
tells the user that there is no warranty for the work (except to the
|
||||
extent that warranties are provided), that licensees may convey the
|
||||
work under this License, and how to view a copy of this License. If
|
||||
the interface presents a list of user commands or options, such as a
|
||||
menu, a prominent item in the list meets this criterion.
|
||||
|
||||
1. Source Code.
|
||||
|
||||
The "source code" for a work means the preferred form of the work
|
||||
for making modifications to it. "Object code" means any non-source
|
||||
form of a work.
|
||||
|
||||
A "Standard Interface" means an interface that either is an official
|
||||
standard defined by a recognized standards body, or, in the case of
|
||||
interfaces specified for a particular programming language, one that
|
||||
is widely used among developers working in that language.
|
||||
|
||||
The "System Libraries" of an executable work include anything, other
|
||||
than the work as a whole, that (a) is included in the normal form of
|
||||
packaging a Major Component, but which is not part of that Major
|
||||
Component, and (b) serves only to enable use of the work with that
|
||||
Major Component, or to implement a Standard Interface for which an
|
||||
implementation is available to the public in source code form. A
|
||||
"Major Component", in this context, means a major essential component
|
||||
(kernel, window system, and so on) of the specific operating system
|
||||
(if any) on which the executable work runs, or a compiler used to
|
||||
produce the work, or an object code interpreter used to run it.
|
||||
|
||||
The "Corresponding Source" for a work in object code form means all
|
||||
the source code needed to generate, install, and (for an executable
|
||||
work) run the object code and to modify the work, including scripts to
|
||||
control those activities. However, it does not include the work's
|
||||
System Libraries, or general-purpose tools or generally available free
|
||||
programs which are used unmodified in performing those activities but
|
||||
which are not part of the work. For example, Corresponding Source
|
||||
includes interface definition files associated with source files for
|
||||
the work, and the source code for shared libraries and dynamically
|
||||
linked subprograms that the work is specifically designed to require,
|
||||
such as by intimate data communication or control flow between those
|
||||
subprograms and other parts of the work.
|
||||
|
||||
The Corresponding Source need not include anything that users
|
||||
can regenerate automatically from other parts of the Corresponding
|
||||
Source.
|
||||
|
||||
The Corresponding Source for a work in source code form is that
|
||||
same work.
|
||||
|
||||
2. Basic Permissions.
|
||||
|
||||
All rights granted under this License are granted for the term of
|
||||
copyright on the Program, and are irrevocable provided the stated
|
||||
conditions are met. This License explicitly affirms your unlimited
|
||||
permission to run the unmodified Program. The output from running a
|
||||
covered work is covered by this License only if the output, given its
|
||||
content, constitutes a covered work. This License acknowledges your
|
||||
rights of fair use or other equivalent, as provided by copyright law.
|
||||
|
||||
You may make, run and propagate covered works that you do not
|
||||
convey, without conditions so long as your license otherwise remains
|
||||
in force. You may convey covered works to others for the sole purpose
|
||||
of having them make modifications exclusively for you, or provide you
|
||||
with facilities for running those works, provided that you comply with
|
||||
the terms of this License in conveying all material for which you do
|
||||
not control copyright. Those thus making or running the covered works
|
||||
for you must do so exclusively on your behalf, under your direction
|
||||
and control, on terms that prohibit them from making any copies of
|
||||
your copyrighted material outside their relationship with you.
|
||||
|
||||
Conveying under any other circumstances is permitted solely under
|
||||
the conditions stated below. Sublicensing is not allowed; section 10
|
||||
makes it unnecessary.
|
||||
|
||||
3. Protecting Users' Legal Rights From Anti-Circumvention Law.
|
||||
|
||||
No covered work shall be deemed part of an effective technological
|
||||
measure under any applicable law fulfilling obligations under article
|
||||
11 of the WIPO copyright treaty adopted on 20 December 1996, or
|
||||
similar laws prohibiting or restricting circumvention of such
|
||||
measures.
|
||||
|
||||
When you convey a covered work, you waive any legal power to forbid
|
||||
circumvention of technological measures to the extent such circumvention
|
||||
is effected by exercising rights under this License with respect to
|
||||
the covered work, and you disclaim any intention to limit operation or
|
||||
modification of the work as a means of enforcing, against the work's
|
||||
users, your or third parties' legal rights to forbid circumvention of
|
||||
technological measures.
|
||||
|
||||
4. Conveying Verbatim Copies.
|
||||
|
||||
You may convey verbatim copies of the Program's source code as you
|
||||
receive it, in any medium, provided that you conspicuously and
|
||||
appropriately publish on each copy an appropriate copyright notice;
|
||||
keep intact all notices stating that this License and any
|
||||
non-permissive terms added in accord with section 7 apply to the code;
|
||||
keep intact all notices of the absence of any warranty; and give all
|
||||
recipients a copy of this License along with the Program.
|
||||
|
||||
You may charge any price or no price for each copy that you convey,
|
||||
and you may offer support or warranty protection for a fee.
|
||||
|
||||
5. Conveying Modified Source Versions.
|
||||
|
||||
You may convey a work based on the Program, or the modifications to
|
||||
produce it from the Program, in the form of source code under the
|
||||
terms of section 4, provided that you also meet all of these conditions:
|
||||
|
||||
a) The work must carry prominent notices stating that you modified
|
||||
it, and giving a relevant date.
|
||||
|
||||
b) The work must carry prominent notices stating that it is
|
||||
released under this License and any conditions added under section
|
||||
7. This requirement modifies the requirement in section 4 to
|
||||
"keep intact all notices".
|
||||
|
||||
c) You must license the entire work, as a whole, under this
|
||||
License to anyone who comes into possession of a copy. This
|
||||
License will therefore apply, along with any applicable section 7
|
||||
additional terms, to the whole of the work, and all its parts,
|
||||
regardless of how they are packaged. This License gives no
|
||||
permission to license the work in any other way, but it does not
|
||||
invalidate such permission if you have separately received it.
|
||||
|
||||
d) If the work has interactive user interfaces, each must display
|
||||
Appropriate Legal Notices; however, if the Program has interactive
|
||||
interfaces that do not display Appropriate Legal Notices, your
|
||||
work need not make them do so.
|
||||
|
||||
A compilation of a covered work with other separate and independent
|
||||
works, which are not by their nature extensions of the covered work,
|
||||
and which are not combined with it such as to form a larger program,
|
||||
in or on a volume of a storage or distribution medium, is called an
|
||||
"aggregate" if the compilation and its resulting copyright are not
|
||||
used to limit the access or legal rights of the compilation's users
|
||||
beyond what the individual works permit. Inclusion of a covered work
|
||||
in an aggregate does not cause this License to apply to the other
|
||||
parts of the aggregate.
|
||||
|
||||
6. Conveying Non-Source Forms.
|
||||
|
||||
You may convey a covered work in object code form under the terms
|
||||
of sections 4 and 5, provided that you also convey the
|
||||
machine-readable Corresponding Source under the terms of this License,
|
||||
in one of these ways:
|
||||
|
||||
a) Convey the object code in, or embodied in, a physical product
|
||||
(including a physical distribution medium), accompanied by the
|
||||
Corresponding Source fixed on a durable physical medium
|
||||
customarily used for software interchange.
|
||||
|
||||
b) Convey the object code in, or embodied in, a physical product
|
||||
(including a physical distribution medium), accompanied by a
|
||||
written offer, valid for at least three years and valid for as
|
||||
long as you offer spare parts or customer support for that product
|
||||
model, to give anyone who possesses the object code either (1) a
|
||||
copy of the Corresponding Source for all the software in the
|
||||
product that is covered by this License, on a durable physical
|
||||
medium customarily used for software interchange, for a price no
|
||||
more than your reasonable cost of physically performing this
|
||||
conveying of source, or (2) access to copy the
|
||||
Corresponding Source from a network server at no charge.
|
||||
|
||||
c) Convey individual copies of the object code with a copy of the
|
||||
written offer to provide the Corresponding Source. This
|
||||
alternative is allowed only occasionally and noncommercially, and
|
||||
only if you received the object code with such an offer, in accord
|
||||
with subsection 6b.
|
||||
|
||||
d) Convey the object code by offering access from a designated
|
||||
place (gratis or for a charge), and offer equivalent access to the
|
||||
Corresponding Source in the same way through the same place at no
|
||||
further charge. You need not require recipients to copy the
|
||||
Corresponding Source along with the object code. If the place to
|
||||
copy the object code is a network server, the Corresponding Source
|
||||
may be on a different server (operated by you or a third party)
|
||||
that supports equivalent copying facilities, provided you maintain
|
||||
clear directions next to the object code saying where to find the
|
||||
Corresponding Source. Regardless of what server hosts the
|
||||
Corresponding Source, you remain obligated to ensure that it is
|
||||
available for as long as needed to satisfy these requirements.
|
||||
|
||||
e) Convey the object code using peer-to-peer transmission, provided
|
||||
you inform other peers where the object code and Corresponding
|
||||
Source of the work are being offered to the general public at no
|
||||
charge under subsection 6d.
|
||||
|
||||
A separable portion of the object code, whose source code is excluded
|
||||
from the Corresponding Source as a System Library, need not be
|
||||
included in conveying the object code work.
|
||||
|
||||
A "User Product" is either (1) a "consumer product", which means any
|
||||
tangible personal property which is normally used for personal, family,
|
||||
or household purposes, or (2) anything designed or sold for incorporation
|
||||
into a dwelling. In determining whether a product is a consumer product,
|
||||
doubtful cases shall be resolved in favor of coverage. For a particular
|
||||
product received by a particular user, "normally used" refers to a
|
||||
typical or common use of that class of product, regardless of the status
|
||||
of the particular user or of the way in which the particular user
|
||||
actually uses, or expects or is expected to use, the product. A product
|
||||
is a consumer product regardless of whether the product has substantial
|
||||
commercial, industrial or non-consumer uses, unless such uses represent
|
||||
the only significant mode of use of the product.
|
||||
|
||||
"Installation Information" for a User Product means any methods,
|
||||
procedures, authorization keys, or other information required to install
|
||||
and execute modified versions of a covered work in that User Product from
|
||||
a modified version of its Corresponding Source. The information must
|
||||
suffice to ensure that the continued functioning of the modified object
|
||||
code is in no case prevented or interfered with solely because
|
||||
modification has been made.
|
||||
|
||||
If you convey an object code work under this section in, or with, or
|
||||
specifically for use in, a User Product, and the conveying occurs as
|
||||
part of a transaction in which the right of possession and use of the
|
||||
User Product is transferred to the recipient in perpetuity or for a
|
||||
fixed term (regardless of how the transaction is characterized), the
|
||||
Corresponding Source conveyed under this section must be accompanied
|
||||
by the Installation Information. But this requirement does not apply
|
||||
if neither you nor any third party retains the ability to install
|
||||
modified object code on the User Product (for example, the work has
|
||||
been installed in ROM).
|
||||
|
||||
The requirement to provide Installation Information does not include a
|
||||
requirement to continue to provide support service, warranty, or updates
|
||||
for a work that has been modified or installed by the recipient, or for
|
||||
the User Product in which it has been modified or installed. Access to a
|
||||
network may be denied when the modification itself materially and
|
||||
adversely affects the operation of the network or violates the rules and
|
||||
protocols for communication across the network.
|
||||
|
||||
Corresponding Source conveyed, and Installation Information provided,
|
||||
in accord with this section must be in a format that is publicly
|
||||
documented (and with an implementation available to the public in
|
||||
source code form), and must require no special password or key for
|
||||
unpacking, reading or copying.
|
||||
|
||||
7. Additional Terms.
|
||||
|
||||
"Additional permissions" are terms that supplement the terms of this
|
||||
License by making exceptions from one or more of its conditions.
|
||||
Additional permissions that are applicable to the entire Program shall
|
||||
be treated as though they were included in this License, to the extent
|
||||
that they are valid under applicable law. If additional permissions
|
||||
apply only to part of the Program, that part may be used separately
|
||||
under those permissions, but the entire Program remains governed by
|
||||
this License without regard to the additional permissions.
|
||||
|
||||
When you convey a copy of a covered work, you may at your option
|
||||
remove any additional permissions from that copy, or from any part of
|
||||
it. (Additional permissions may be written to require their own
|
||||
removal in certain cases when you modify the work.) You may place
|
||||
additional permissions on material, added by you to a covered work,
|
||||
for which you have or can give appropriate copyright permission.
|
||||
|
||||
Notwithstanding any other provision of this License, for material you
|
||||
add to a covered work, you may (if authorized by the copyright holders of
|
||||
that material) supplement the terms of this License with terms:
|
||||
|
||||
a) Disclaiming warranty or limiting liability differently from the
|
||||
terms of sections 15 and 16 of this License; or
|
||||
|
||||
b) Requiring preservation of specified reasonable legal notices or
|
||||
author attributions in that material or in the Appropriate Legal
|
||||
Notices displayed by works containing it; or
|
||||
|
||||
c) Prohibiting misrepresentation of the origin of that material, or
|
||||
requiring that modified versions of such material be marked in
|
||||
reasonable ways as different from the original version; or
|
||||
|
||||
d) Limiting the use for publicity purposes of names of licensors or
|
||||
authors of the material; or
|
||||
|
||||
e) Declining to grant rights under trademark law for use of some
|
||||
trade names, trademarks, or service marks; or
|
||||
|
||||
f) Requiring indemnification of licensors and authors of that
|
||||
material by anyone who conveys the material (or modified versions of
|
||||
it) with contractual assumptions of liability to the recipient, for
|
||||
any liability that these contractual assumptions directly impose on
|
||||
those licensors and authors.
|
||||
|
||||
All other non-permissive additional terms are considered "further
|
||||
restrictions" within the meaning of section 10. If the Program as you
|
||||
received it, or any part of it, contains a notice stating that it is
|
||||
governed by this License along with a term that is a further
|
||||
restriction, you may remove that term. If a license document contains
|
||||
a further restriction but permits relicensing or conveying under this
|
||||
License, you may add to a covered work material governed by the terms
|
||||
of that license document, provided that the further restriction does
|
||||
not survive such relicensing or conveying.
|
||||
|
||||
If you add terms to a covered work in accord with this section, you
|
||||
must place, in the relevant source files, a statement of the
|
||||
additional terms that apply to those files, or a notice indicating
|
||||
where to find the applicable terms.
|
||||
|
||||
Additional terms, permissive or non-permissive, may be stated in the
|
||||
form of a separately written license, or stated as exceptions;
|
||||
the above requirements apply either way.
|
||||
|
||||
8. Termination.
|
||||
|
||||
You may not propagate or modify a covered work except as expressly
|
||||
provided under this License. Any attempt otherwise to propagate or
|
||||
modify it is void, and will automatically terminate your rights under
|
||||
this License (including any patent licenses granted under the third
|
||||
paragraph of section 11).
|
||||
|
||||
However, if you cease all violation of this License, then your
|
||||
license from a particular copyright holder is reinstated (a)
|
||||
provisionally, unless and until the copyright holder explicitly and
|
||||
finally terminates your license, and (b) permanently, if the copyright
|
||||
holder fails to notify you of the violation by some reasonable means
|
||||
prior to 60 days after the cessation.
|
||||
|
||||
Moreover, your license from a particular copyright holder is
|
||||
reinstated permanently if the copyright holder notifies you of the
|
||||
violation by some reasonable means, this is the first time you have
|
||||
received notice of violation of this License (for any work) from that
|
||||
copyright holder, and you cure the violation prior to 30 days after
|
||||
your receipt of the notice.
|
||||
|
||||
Termination of your rights under this section does not terminate the
|
||||
licenses of parties who have received copies or rights from you under
|
||||
this License. If your rights have been terminated and not permanently
|
||||
reinstated, you do not qualify to receive new licenses for the same
|
||||
material under section 10.
|
||||
|
||||
9. Acceptance Not Required for Having Copies.
|
||||
|
||||
You are not required to accept this License in order to receive or
|
||||
run a copy of the Program. Ancillary propagation of a covered work
|
||||
occurring solely as a consequence of using peer-to-peer transmission
|
||||
to receive a copy likewise does not require acceptance. However,
|
||||
nothing other than this License grants you permission to propagate or
|
||||
modify any covered work. These actions infringe copyright if you do
|
||||
not accept this License. Therefore, by modifying or propagating a
|
||||
covered work, you indicate your acceptance of this License to do so.
|
||||
|
||||
10. Automatic Licensing of Downstream Recipients.
|
||||
|
||||
Each time you convey a covered work, the recipient automatically
|
||||
receives a license from the original licensors, to run, modify and
|
||||
propagate that work, subject to this License. You are not responsible
|
||||
for enforcing compliance by third parties with this License.
|
||||
|
||||
An "entity transaction" is a transaction transferring control of an
|
||||
organization, or substantially all assets of one, or subdividing an
|
||||
organization, or merging organizations. If propagation of a covered
|
||||
work results from an entity transaction, each party to that
|
||||
transaction who receives a copy of the work also receives whatever
|
||||
licenses to the work the party's predecessor in interest had or could
|
||||
give under the previous paragraph, plus a right to possession of the
|
||||
Corresponding Source of the work from the predecessor in interest, if
|
||||
the predecessor has it or can get it with reasonable efforts.
|
||||
|
||||
You may not impose any further restrictions on the exercise of the
|
||||
rights granted or affirmed under this License. For example, you may
|
||||
not impose a license fee, royalty, or other charge for exercise of
|
||||
rights granted under this License, and you may not initiate litigation
|
||||
(including a cross-claim or counterclaim in a lawsuit) alleging that
|
||||
any patent claim is infringed by making, using, selling, offering for
|
||||
sale, or importing the Program or any portion of it.
|
||||
|
||||
11. Patents.
|
||||
|
||||
A "contributor" is a copyright holder who authorizes use under this
|
||||
License of the Program or a work on which the Program is based. The
|
||||
work thus licensed is called the contributor's "contributor version".
|
||||
|
||||
A contributor's "essential patent claims" are all patent claims
|
||||
owned or controlled by the contributor, whether already acquired or
|
||||
hereafter acquired, that would be infringed by some manner, permitted
|
||||
by this License, of making, using, or selling its contributor version,
|
||||
but do not include claims that would be infringed only as a
|
||||
consequence of further modification of the contributor version. For
|
||||
purposes of this definition, "control" includes the right to grant
|
||||
patent sublicenses in a manner consistent with the requirements of
|
||||
this License.
|
||||
|
||||
Each contributor grants you a non-exclusive, worldwide, royalty-free
|
||||
patent license under the contributor's essential patent claims, to
|
||||
make, use, sell, offer for sale, import and otherwise run, modify and
|
||||
propagate the contents of its contributor version.
|
||||
|
||||
In the following three paragraphs, a "patent license" is any express
|
||||
agreement or commitment, however denominated, not to enforce a patent
|
||||
(such as an express permission to practice a patent or covenant not to
|
||||
sue for patent infringement). To "grant" such a patent license to a
|
||||
party means to make such an agreement or commitment not to enforce a
|
||||
patent against the party.
|
||||
|
||||
If you convey a covered work, knowingly relying on a patent license,
|
||||
and the Corresponding Source of the work is not available for anyone
|
||||
to copy, free of charge and under the terms of this License, through a
|
||||
publicly available network server or other readily accessible means,
|
||||
then you must either (1) cause the Corresponding Source to be so
|
||||
available, or (2) arrange to deprive yourself of the benefit of the
|
||||
patent license for this particular work, or (3) arrange, in a manner
|
||||
consistent with the requirements of this License, to extend the patent
|
||||
license to downstream recipients. "Knowingly relying" means you have
|
||||
actual knowledge that, but for the patent license, your conveying the
|
||||
covered work in a country, or your recipient's use of the covered work
|
||||
in a country, would infringe one or more identifiable patents in that
|
||||
country that you have reason to believe are valid.
|
||||
|
||||
If, pursuant to or in connection with a single transaction or
|
||||
arrangement, you convey, or propagate by procuring conveyance of, a
|
||||
covered work, and grant a patent license to some of the parties
|
||||
receiving the covered work authorizing them to use, propagate, modify
|
||||
or convey a specific copy of the covered work, then the patent license
|
||||
you grant is automatically extended to all recipients of the covered
|
||||
work and works based on it.
|
||||
|
||||
A patent license is "discriminatory" if it does not include within
|
||||
the scope of its coverage, prohibits the exercise of, or is
|
||||
conditioned on the non-exercise of one or more of the rights that are
|
||||
specifically granted under this License. You may not convey a covered
|
||||
work if you are a party to an arrangement with a third party that is
|
||||
in the business of distributing software, under which you make payment
|
||||
to the third party based on the extent of your activity of conveying
|
||||
the work, and under which the third party grants, to any of the
|
||||
parties who would receive the covered work from you, a discriminatory
|
||||
patent license (a) in connection with copies of the covered work
|
||||
conveyed by you (or copies made from those copies), or (b) primarily
|
||||
for and in connection with specific products or compilations that
|
||||
contain the covered work, unless you entered into that arrangement,
|
||||
or that patent license was granted, prior to 28 March 2007.
|
||||
|
||||
Nothing in this License shall be construed as excluding or limiting
|
||||
any implied license or other defenses to infringement that may
|
||||
otherwise be available to you under applicable patent law.
|
||||
|
||||
12. No Surrender of Others' Freedom.
|
||||
|
||||
If conditions are imposed on you (whether by court order, agreement or
|
||||
otherwise) that contradict the conditions of this License, they do not
|
||||
excuse you from the conditions of this License. If you cannot convey a
|
||||
covered work so as to satisfy simultaneously your obligations under this
|
||||
License and any other pertinent obligations, then as a consequence you may
|
||||
not convey it at all. For example, if you agree to terms that obligate you
|
||||
to collect a royalty for further conveying from those to whom you convey
|
||||
the Program, the only way you could satisfy both those terms and this
|
||||
License would be to refrain entirely from conveying the Program.
|
||||
|
||||
13. Remote Network Interaction; Use with the GNU General Public License.
|
||||
|
||||
Notwithstanding any other provision of this License, if you modify the
|
||||
Program, your modified version must prominently offer all users
|
||||
interacting with it remotely through a computer network (if your version
|
||||
supports such interaction) an opportunity to receive the Corresponding
|
||||
Source of your version by providing access to the Corresponding Source
|
||||
from a network server at no charge, through some standard or customary
|
||||
means of facilitating copying of software. This Corresponding Source
|
||||
shall include the Corresponding Source for any work covered by version 3
|
||||
of the GNU General Public License that is incorporated pursuant to the
|
||||
following paragraph.
|
||||
|
||||
Notwithstanding any other provision of this License, you have
|
||||
permission to link or combine any covered work with a work licensed
|
||||
under version 3 of the GNU General Public License into a single
|
||||
combined work, and to convey the resulting work. The terms of this
|
||||
License will continue to apply to the part which is the covered work,
|
||||
but the work with which it is combined will remain governed by version
|
||||
3 of the GNU General Public License.
|
||||
|
||||
14. Revised Versions of this License.
|
||||
|
||||
The Free Software Foundation may publish revised and/or new versions of
|
||||
the GNU Affero General Public License from time to time. Such new versions
|
||||
will be similar in spirit to the present version, but may differ in detail to
|
||||
address new problems or concerns.
|
||||
|
||||
Each version is given a distinguishing version number. If the
|
||||
Program specifies that a certain numbered version of the GNU Affero General
|
||||
Public License "or any later version" applies to it, you have the
|
||||
option of following the terms and conditions either of that numbered
|
||||
version or of any later version published by the Free Software
|
||||
Foundation. If the Program does not specify a version number of the
|
||||
GNU Affero General Public License, you may choose any version ever published
|
||||
by the Free Software Foundation.
|
||||
|
||||
If the Program specifies that a proxy can decide which future
|
||||
versions of the GNU Affero General Public License can be used, that proxy's
|
||||
public statement of acceptance of a version permanently authorizes you
|
||||
to choose that version for the Program.
|
||||
|
||||
Later license versions may give you additional or different
|
||||
permissions. However, no additional obligations are imposed on any
|
||||
author or copyright holder as a result of your choosing to follow a
|
||||
later version.
|
||||
|
||||
15. Disclaimer of Warranty.
|
||||
|
||||
THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
|
||||
APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
|
||||
HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
|
||||
OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
|
||||
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
|
||||
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
|
||||
IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
|
||||
ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
|
||||
|
||||
16. Limitation of Liability.
|
||||
|
||||
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
|
||||
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
|
||||
THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
|
||||
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
|
||||
USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
|
||||
DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
|
||||
PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
|
||||
EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
|
||||
SUCH DAMAGES.
|
||||
|
||||
17. Interpretation of Sections 15 and 16.
|
||||
|
||||
If the disclaimer of warranty and limitation of liability provided
|
||||
above cannot be given local legal effect according to their terms,
|
||||
reviewing courts shall apply local law that most closely approximates
|
||||
an absolute waiver of all civil liability in connection with the
|
||||
Program, unless a warranty or assumption of liability accompanies a
|
||||
copy of the Program in return for a fee.
|
||||
|
||||
END OF TERMS AND CONDITIONS
|
||||
|
||||
How to Apply These Terms to Your New Programs
|
||||
|
||||
If you develop a new program, and you want it to be of the greatest
|
||||
possible use to the public, the best way to achieve this is to make it
|
||||
free software which everyone can redistribute and change under these terms.
|
||||
|
||||
To do so, attach the following notices to the program. It is safest
|
||||
to attach them to the start of each source file to most effectively
|
||||
state the exclusion of warranty; and each file should have at least
|
||||
the "copyright" line and a pointer to where the full notice is found.
|
||||
|
||||
<one line to give the program's name and a brief idea of what it does.>
|
||||
Copyright (C) <year> <name of author>
|
||||
|
||||
This program is free software: you can redistribute it and/or modify
|
||||
it under the terms of the GNU Affero General Public License as published
|
||||
by the Free Software Foundation, either version 3 of the License, or
|
||||
(at your option) any later version.
|
||||
|
||||
This program is distributed in the hope that it will be useful,
|
||||
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||||
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||||
GNU Affero General Public License for more details.
|
||||
|
||||
You should have received a copy of the GNU Affero General Public License
|
||||
along with this program. If not, see <https://www.gnu.org/licenses/>.
|
||||
|
||||
Also add information on how to contact you by electronic and paper mail.
|
||||
|
||||
If your software can interact with users remotely through a computer
|
||||
network, you should also make sure that it provides a way for users to
|
||||
get its source. For example, if your program is a web application, its
|
||||
interface could display a "Source" link that leads users to an archive
|
||||
of the code. There are many ways you could offer source, and different
|
||||
solutions will be better for different programs; see section 13 for the
|
||||
specific requirements.
|
||||
|
||||
You should also get your employer (if you work as a programmer) or school,
|
||||
if any, to sign a "copyright disclaimer" for the program, if necessary.
|
||||
For more information on this, and how to apply and follow the GNU AGPL, see
|
||||
<https://www.gnu.org/licenses/>.
|
||||
@ -0,0 +1,836 @@
|
||||
import datetime
|
||||
import os
|
||||
import pickle
|
||||
from collections import Counter
|
||||
from utils import get_new_cols
|
||||
import numpy as np
|
||||
import pandas as pd
|
||||
from tqdm import tqdm
|
||||
from gensim.models import Word2Vec
|
||||
from utils import GENERATION_DIR
|
||||
from utils import KEY_1, KEY_2, KEY_3, KEY_4
|
||||
from tqdm import tqdm
|
||||
from scipy import stats
|
||||
|
||||
def cat2num(df, cat_cols, Transfer2num=True):
|
||||
'''
|
||||
|
||||
:param df:
|
||||
:param cat_cols: 类别特征列表
|
||||
:param Transfer2num: 类别特征转换为数值特征
|
||||
:return:
|
||||
'''
|
||||
if Transfer2num:
|
||||
|
||||
print('Transfer category feature to num feature ')
|
||||
for col in cat_cols:
|
||||
|
||||
if not os.path.exists(os.path.join(GENERATION_DIR, f'{col}_map.pkl')):
|
||||
print(f'Transfer : {col}')
|
||||
tmp_map = dict(zip(df[col].unique(), range(df[col].nunique())))
|
||||
with open(os.path.join(GENERATION_DIR, f'{col}_map.pkl'), 'wb') as f:
|
||||
pickle.dump(tmp_map, f)
|
||||
else:
|
||||
with open(os.path.join(GENERATION_DIR, f'{col}_map.pkl'), 'rb') as f:
|
||||
tmp_map = pickle.load(f)
|
||||
df[f'{col}_LabelEnc'] = df[col].map(tmp_map).fillna(-1).astype(int)
|
||||
else:
|
||||
print('Transfer category feature to category feature ')
|
||||
for col in cat_cols:
|
||||
df[col] = df[col].astype('category')
|
||||
print('Transfer category feature to num feature Down...')
|
||||
return df
|
||||
|
||||
def add_minutes(x, minutes=5):
|
||||
dt = datetime.datetime.strptime(x, '%Y-%m-%d %H:%M:%S')
|
||||
out_date = (dt + datetime.timedelta(minutes=minutes)
|
||||
).strftime('%Y-%m-%d %H:%M:%S')
|
||||
return out_date
|
||||
|
||||
|
||||
def time_process(df, time_cols, minutes_):
|
||||
df[f'time_{minutes_}'] = df[time_cols].apply(
|
||||
lambda x: add_minutes(str(x), minutes_))
|
||||
return df
|
||||
|
||||
|
||||
def get_fea(x, fea):
|
||||
if fea in x:
|
||||
return 1
|
||||
else:
|
||||
return 0
|
||||
|
||||
|
||||
def get_last_msg_cnt(x):
|
||||
last_msg = x[-1]
|
||||
cnt = x.count(last_msg)
|
||||
return cnt
|
||||
|
||||
|
||||
def get_first_msg_cnt(x):
|
||||
first_msg = x[0]
|
||||
cnt = x.count(first_msg)
|
||||
return cnt
|
||||
|
||||
|
||||
def add_last_next_time4fault(label, preliminary_submit_dataset_a,
|
||||
time_interval, next_time_list):
|
||||
print(f'添加自定义异常出现的时间间隔{time_interval}的前后的时间点')
|
||||
for i in tqdm([-i for i in next_time_list] + next_time_list):
|
||||
label = time_process(label, 'fault_time', i * time_interval)
|
||||
preliminary_submit_dataset_a = time_process(
|
||||
preliminary_submit_dataset_a, 'fault_time', i * time_interval)
|
||||
|
||||
return label, preliminary_submit_dataset_a
|
||||
|
||||
|
||||
def get_msg_text_fea(df, msg_type='last'):
|
||||
print(f'获取 msg text {msg_type}特征')
|
||||
|
||||
df_fea = df.groupby(['sn', 'fault_time']).agg(
|
||||
{'msg_list': 'sum', 'msg_0': 'sum', 'msg_1': 'sum', 'msg_2': 'sum'}).reset_index()
|
||||
df_fea['msg_list_unique'] = df_fea['msg_list'].apply(lambda x: str(set(x)))
|
||||
df_fea['msg_0_unique'] = df_fea['msg_0'].apply(lambda x: str(set(x)))
|
||||
df_fea['msg_1_unique'] = df_fea['msg_1'].apply(lambda x: str(set(x)))
|
||||
df_fea['msg_2_unique'] = df_fea['msg_2'].apply(lambda x: str(set(x)))
|
||||
|
||||
df_fea['msg_list_list'] = df_fea['msg_list'].apply(lambda x: str(x))
|
||||
df_fea['msg_0_list'] = df_fea['msg_0'].apply(lambda x: str(x))
|
||||
df_fea['msg_1_list'] = df_fea['msg_1'].apply(lambda x: str(x))
|
||||
df_fea['msg_2_list'] = df_fea['msg_2'].apply(lambda x: str(x))
|
||||
|
||||
df_fea['msg_0_first'] = df_fea['msg_0'].apply(lambda x: x[0])
|
||||
df_fea['msg_1_first'] = df_fea['msg_1'].apply(lambda x: x[0])
|
||||
df_fea['msg_2_first'] = df_fea['msg_2'].apply(lambda x: x[0])
|
||||
|
||||
df_fea['msg_0_last'] = df_fea['msg_0'].apply(lambda x: x[-1])
|
||||
df_fea['msg_1_last'] = df_fea['msg_1'].apply(lambda x: x[-1])
|
||||
df_fea['msg_2_last'] = df_fea['msg_2'].apply(lambda x: x[-1])
|
||||
|
||||
df_fea['msg_last'] = df.groupby(['sn', 'fault_time']).apply(
|
||||
lambda x: x['msg'].to_list()[-1]).values
|
||||
df_fea['msg_first'] = df.groupby(['sn', 'fault_time']).apply(
|
||||
lambda x: x['msg'].to_list()[0]).values
|
||||
|
||||
df_fea['last_msg_cnt'] = df_fea['msg_list'].apply(
|
||||
lambda x: get_last_msg_cnt(x))
|
||||
df_fea['first_msg_cnt'] = df_fea['msg_list'].apply(
|
||||
lambda x: get_first_msg_cnt(x))
|
||||
cat_cols = ['msg_list', 'msg_0', 'msg_1', 'msg_2',
|
||||
'msg_list_unique', 'msg_0_unique', 'msg_1_unique', 'msg_2_unique',
|
||||
'msg_list_list', 'msg_0_list', 'msg_1_list', 'msg_2_list',
|
||||
'msg_0_first', 'msg_1_first', 'msg_2_first', 'msg_0_last', 'msg_1_last',
|
||||
'msg_2_last', 'msg_last', 'msg_first']
|
||||
num_cols = ['last_msg_cnt', 'first_msg_cnt']
|
||||
id_cols = ['sn', 'fault_time']
|
||||
|
||||
df_fea = df_fea.rename(
|
||||
columns={
|
||||
i: f'{msg_type}_{i}' for i in (cat_cols + num_cols)})
|
||||
cat_cols = [f'{msg_type}_{i}' for i in cat_cols]
|
||||
for cat_col in cat_cols:
|
||||
df_fea[cat_col] = df_fea[cat_col].astype(str)
|
||||
df_fea = cat2num(df_fea, cat_cols, Transfer2num=True)
|
||||
for i in cat_cols:
|
||||
del df_fea[i]
|
||||
return df_fea
|
||||
|
||||
def add_w2v_feats(all_data,w2v_feats_df,f1,f2,emb_size = 32,window = 5,min_count =5,):
|
||||
print(f'生成 {f1}_{f2}_w2v 特征')
|
||||
|
||||
df_fea = all_data.groupby(f1).agg({f2:'sum'}).reset_index()
|
||||
df_emb = df_fea[[f1 ]]
|
||||
sencences = df_fea[f2].to_list()
|
||||
if not os.path.exists(os.path.join(GENERATION_DIR, f'{f1}_{f2}_w2v_model.pkl')):
|
||||
print(f'{f1}_{f2}_w2v_model 不存在，开始训练......')
|
||||
model = Word2Vec(sencences, vector_size=emb_size, window=window,
|
||||
min_count=min_count, sg=0, hs=1, seed=42)
|
||||
with open(os.path.join(GENERATION_DIR, f'{f1}_{f2}_w2v_model.pkl'), 'wb') as f:
|
||||
pickle.dump(model, f)
|
||||
else:
|
||||
print(f'{f1}_{f2}_w2v_model 已存在，开始读取......')
|
||||
with open(os.path.join(GENERATION_DIR, f'{f1}_{f2}_w2v_model.pkl'), 'rb') as f:
|
||||
model = pickle.load(f)
|
||||
|
||||
emb_matrix_mean = []
|
||||
for sent in sencences:
|
||||
vec = []
|
||||
for w in sent:
|
||||
if w in model.wv:
|
||||
vec.append(model.wv[w])
|
||||
if len(vec) >0:
|
||||
emb_matrix_mean.append(np.mean(vec,axis = 0))
|
||||
else:
|
||||
emb_matrix_mean.append([0]*emb_size)
|
||||
df_emb_mean = pd.DataFrame(emb_matrix_mean).add_prefix(f'{f1}_{f2}_w2v_')
|
||||
|
||||
df_emb = pd.concat([df_emb,df_emb_mean],axis = 1)
|
||||
w2v_feats_df = w2v_feats_df.merge(df_emb,on = f1,how ='left')
|
||||
return w2v_feats_df
|
||||
def get_w2v_feats(all_data,f1_list,f2_list):
|
||||
all_data['msg_list'] = all_data['msg'].apply(lambda x: [i.strip() for i in x.split(' | ')])
|
||||
all_data['msg_0'] = all_data['msg'].apply(lambda x: [get_msg_location(x.split(' | '), 0)])
|
||||
all_data['msg_1'] = all_data['msg'].apply(lambda x: [get_msg_location(x.split(' | '), 1)])
|
||||
all_data['msg_2'] = all_data['msg'].apply(lambda x: [get_msg_location(x.split(' | '), 2)])
|
||||
w2v_feats_df = all_data[f1_list].drop_duplicates()
|
||||
for f1 in f1_list:
|
||||
for f2 in f2_list:
|
||||
w2v_feats_df = add_w2v_feats(all_data,w2v_feats_df,f1,f2,emb_size = 10,window = 5,min_count =5,)
|
||||
print(f'w2v_feats 的特征维度: {w2v_feats_df.shape}')
|
||||
return w2v_feats_df
|
||||
|
||||
|
||||
|
||||
def get_time_diff_feats_v2(all_data):
|
||||
print('生成时间差特征 time_diff_feats_v2')
|
||||
all_data['duration_seconds'] = all_data['time_interval']
|
||||
all_data['duration_minutes'] = all_data['time_interval'] / 60
|
||||
df_merge_log = all_data[['sn', 'fault_time', 'label', 'time', 'msg',
|
||||
'server_model', 'time_interval', 'duration_seconds',
|
||||
'duration_minutes']]
|
||||
df_merge_log['fault_id'] = df_merge_log['sn'] + '_' + df_merge_log['fault_time'] + '_' + df_merge_log[
|
||||
'server_model']
|
||||
f1_list = ['fault_id', 'sn', 'server_model']
|
||||
f2_list = ['duration_minutes', 'duration_seconds']
|
||||
time_diff_feats_v2 = df_merge_log[['sn', 'fault_time', 'fault_id', 'server_model']].drop_duplicates().reset_index(
|
||||
drop=True)
|
||||
|
||||
for f1 in f1_list:
|
||||
for f2 in f2_list:
|
||||
func_opt = ['count', 'nunique', 'min', 'max', 'median', 'sum']
|
||||
for opt in func_opt:
|
||||
tmp = df_merge_log.groupby([f1])[f2].agg([(f'{f2}_in_{f1}_' + opt, opt)]).reset_index()
|
||||
# print(f'{f1}_in_{f2}_{opt}:{tmp.shape}' )
|
||||
time_diff_feats_v2 = time_diff_feats_v2.merge(tmp, on=f1, how='left')
|
||||
|
||||
temp = df_merge_log.groupby([f1])[f2].apply(lambda x: stats.mode(x)[0][0])
|
||||
time_diff_feats_v2[f'{f2}_in_{f1}_mode'] = time_diff_feats_v2[f1].map(temp).fillna(np.nan)
|
||||
secs = [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9]
|
||||
for sec in secs:
|
||||
temp = df_merge_log.groupby([f1])[f2].quantile(sec).reset_index(
|
||||
name=f'log_{f2}_in_{f1}_quantile_' + str(sec * 100))
|
||||
# print(f'log_{f1}_in_{f2}_quantile_{str(sec * 100)}:{tmp.shape}' )
|
||||
time_diff_feats_v2 = pd.merge(time_diff_feats_v2, temp, on=f1, how='left')
|
||||
del time_diff_feats_v2['fault_id']
|
||||
return time_diff_feats_v2
|
||||
|
||||
def get_feature(data, time_list, log_fea, fea_num, key):
|
||||
print(f'当前特征维度{data.shape}')
|
||||
fea_df_list = []
|
||||
fea_cnt_list = ['OEM record c2', 'Processor CPU_Core_Error', '001c4c', 'System Event Sys_Event','OEM CPU0 MCERR',
|
||||
'OEM CPU0 CATERR', 'Reading 0 < Threshold 2 degrees C', '0203c0a80101',
|
||||
'Unknown CPU0 MCERR', 'Unknown CPU0 CATERR','Memory', 'Correctable ECC logging limit reached',
|
||||
'Memory MEM_CHE0_Status', 'Memory Memory_Status', 'Memory #0x87', 'Memory CPU0F0_DIMM_Stat',
|
||||
'Drive Fault', 'NMI/Diag Interrupt', 'Failure detected', 'Power Supply AC lost', ]
|
||||
for time_tmp in tqdm(time_list):
|
||||
print(f'获取异常前后 {time_tmp} min的数据进行聚合')
|
||||
tmp1 = data[(pd.to_datetime(data['time']) < pd.to_datetime(data[f'time_{time_tmp}'])) & (pd.to_datetime(data['time']) > pd.to_datetime(data[f'time_-{time_tmp}']))].sort_values(
|
||||
['sn', 'fault_time'])
|
||||
tmp1 = tmp1.groupby(key).apply(
|
||||
lambda x: ' | '.join(x['msg'].to_list())).reset_index().rename(columns={0: 'msg'})
|
||||
tmp1[f'msg_len'] = tmp1['msg'].apply(lambda x: len(x.split(' | ')))
|
||||
# tmp1[f'msg_len_two'] = tmp1['msg'].apply(lambda x: len(x))
|
||||
# 添加数字个数
|
||||
# tmp1[f'msg_num_two'] = tmp1['msg'].apply(
|
||||
# lambda x: len([int(s) for s in re.findall(r'\b\d+\b', x)]))
|
||||
print(f'根据异常前后 {time_tmp} min的数据的日志数据提取 {fea_num} 个稀疏特征')
|
||||
feature = log_fea + ['msg_len']
|
||||
for fea in feature:
|
||||
tmp1[fea] = tmp1['msg'].apply(lambda x: get_fea(x, fea))
|
||||
# 添加计数特征
|
||||
if fea in fea_cnt_list:
|
||||
tmp1[f'{fea}_cnt'] = tmp1['msg'].apply(lambda x:x.replace('|',' ').replace('_',' ').split(' ').count(fea))
|
||||
feature.append(f'{fea}_cnt')
|
||||
tmp1_new_col_map = {i: i + '_' + str(int(time_tmp)) for i in feature}
|
||||
tmp1 = tmp1.rename(columns=tmp1_new_col_map)
|
||||
del tmp1['msg']
|
||||
fea_df_list.append(tmp1)
|
||||
fea_df = fea_df_list[-1]
|
||||
print(fea_df.shape)
|
||||
for i in fea_df_list[:-1]:
|
||||
fea_df = fea_df.merge(i, on=key, how='left')
|
||||
print(fea_df.shape)
|
||||
return fea_df
|
||||
|
||||
|
||||
def get_msg_location(x, num):
|
||||
try:
|
||||
return x[num]
|
||||
except BaseException:
|
||||
return '其它'
|
||||
|
||||
|
||||
def get_nearest_msg_fea(train, test):
|
||||
print('生成 nearest_msg 特征')
|
||||
df = pd.concat([train, test], axis=0, ignore_index=True)
|
||||
df['duration_minutes'] = (pd.to_datetime(df['fault_time']) - pd.to_datetime(df['time'])).apply(
|
||||
lambda x: x.total_seconds())
|
||||
df = df.sort_values(
|
||||
['sn', 'server_model', 'fault_time', 'time']).reset_index(drop=True)
|
||||
df['duration_minutes_abs'] = np.abs(df['duration_minutes'])
|
||||
|
||||
df['duration_minutes_abs_rank'] = df.groupby(['sn', 'server_model', 'fault_time'])['duration_minutes_abs'].rank(
|
||||
method='first', ascending=False)
|
||||
|
||||
key = ['sn', 'server_model', 'fault_time', 'duration_minutes_abs']
|
||||
df = df.sort_values(key, ascending=False)
|
||||
df = df.drop_duplicates(
|
||||
['sn', 'server_model', 'fault_time', ], keep='first')
|
||||
|
||||
df.loc[df['duration_minutes'] ==
|
||||
df['duration_minutes_abs'], 'last_or_next'] = 1
|
||||
df.loc[df['duration_minutes'] !=
|
||||
df['duration_minutes_abs'], 'last_or_next'] = 0
|
||||
df['msg_cnt'] = df['msg'].map(df['msg'].value_counts())
|
||||
df['msg_0'] = df['msg'].apply(
|
||||
lambda x: get_msg_location(
|
||||
x.split(' | '), 0))
|
||||
df['msg_0_cnt'] = df['msg_0'].map(df['msg_0'].value_counts())
|
||||
df['msg_1'] = df['msg'].apply(
|
||||
lambda x: get_msg_location(
|
||||
x.split(' | '), 1))
|
||||
df['msg_1_cnt'] = df['msg_1'].map(df['msg_1'].value_counts())
|
||||
df['msg_2'] = df['msg'].apply(
|
||||
lambda x: get_msg_location(
|
||||
x.split(' | '), 2))
|
||||
df['msg_2_cnt'] = df['msg_2'].map(df['msg_2'].value_counts())
|
||||
cat_feats = ['msg', 'msg_0', 'msg_1',
|
||||
'msg_2'] # ,'server_model_day_date','server_model_dayofmonth','server_model_dayofweek','server_model_hour']
|
||||
# for name in cat_feats:
|
||||
# le = LabelEncoder()
|
||||
# df[f'{name}_LabelEnc'] = le.fit_transform(df[name])
|
||||
df = cat2num(df,cat_feats)
|
||||
df = df.drop_duplicates().reset_index(drop=True)
|
||||
df = df[['sn', 'server_model', 'fault_time', 'msg_cnt',
|
||||
'msg_0_cnt', 'msg_1_cnt', 'msg_2_cnt',
|
||||
# 'duration_minutes_abs','duration_minutes', 'duration_minutes_abs_rank',
|
||||
'last_or_next', 'msg_LabelEnc', 'msg_0_LabelEnc', 'msg_1_LabelEnc', 'msg_2_LabelEnc']]
|
||||
print(f'生成 nearest_msg 特征完毕，特征维度{df.shape}')
|
||||
return df
|
||||
|
||||
def get_server_model_time_interval_stat_fea(all_data):
|
||||
server_model_time_interval_stat_fea = all_data.groupby('server_model').agg({'time_interval':['min','max','mean','median']}).reset_index()
|
||||
server_model_time_interval_stat_fea = get_new_cols(server_model_time_interval_stat_fea,key = ['server_model' ])
|
||||
|
||||
server_model_time_interval_stat_fea.columns = ['server_model', 'sm_time_interval_min', 'sm_ttime_interval_max',
|
||||
'sm_ttime_interval_mean', 'sm_ttime_interval_median']
|
||||
return server_model_time_interval_stat_fea
|
||||
|
||||
def get_server_model_sn_fea_2(train, test):
|
||||
df = pd.concat([train[['sn', 'server_model']],
|
||||
test[['sn', 'server_model']]], ignore_index=True)
|
||||
df['server_model_count_sn_2'] = df.groupby(
|
||||
['server_model'])['sn'].transform('count')
|
||||
df['server_model_nunique_sn_2'] = df.groupby(
|
||||
['server_model'])['sn'].transform('nunique')
|
||||
df['sn_cnt_2'] = df['sn'].map(df['sn'].value_counts())
|
||||
return df.drop_duplicates().reset_index(drop=True)
|
||||
|
||||
|
||||
def get_4_time_stat_fea(df):
|
||||
print(' 生成时间统计特征')
|
||||
time_stat_fea_df = df.groupby(['sn', 'fault_time', 'server_model']).agg(
|
||||
{'duration_minutes': ['min', 'max', 'mean', 'median', 'skew', 'sum', 'std', 'count'],
|
||||
'log_duration_minutes': ['min', 'max', 'mean', 'median', 'skew', 'sum', 'std'],
|
||||
'time_diff_1': ['min', 'max', 'mean', 'median', 'skew', 'sum', 'std'],
|
||||
'log_time_diff_1': ['min', 'max', 'median'],
|
||||
}).reset_index()
|
||||
new_time_stat_cols = []
|
||||
for i in time_stat_fea_df.columns:
|
||||
if i[0] in ['sn', 'fault_time', 'server_model']:
|
||||
new_time_stat_cols.append(i[0])
|
||||
else:
|
||||
new_time_stat_cols.append(f'{i[0]}_{i[1]}')
|
||||
# print(f'{i[0]}_{i[1]}')
|
||||
time_stat_fea_df.loc[time_stat_fea_df[i[0]]
|
||||
[i[1]] == -np.inf, (i[0], i[1])] = -20
|
||||
time_stat_fea_df.loc[time_stat_fea_df[i[0]]
|
||||
[i[1]] == np.inf, (i[0], i[1])] = 30
|
||||
time_stat_fea_df.columns = new_time_stat_cols
|
||||
time_stat_fea_df['duration_minutes_range'] = time_stat_fea_df['duration_minutes_max'] - time_stat_fea_df[
|
||||
'duration_minutes_min']
|
||||
time_stat_fea_df['log_duration_minutes_range'] = time_stat_fea_df['log_duration_minutes_max'] - time_stat_fea_df[
|
||||
'log_duration_minutes_min']
|
||||
time_stat_fea_df['time_diff_1_range'] = time_stat_fea_df['time_diff_1_max'] - \
|
||||
time_stat_fea_df['time_diff_1_min']
|
||||
time_stat_fea_df['log_time_diff_1_range'] = time_stat_fea_df['log_time_diff_1_max'] - time_stat_fea_df[
|
||||
'log_time_diff_1_min']
|
||||
time_stat_fea_df['duration_minutes_freq'] = time_stat_fea_df['duration_minutes_range'] / time_stat_fea_df[
|
||||
'duration_minutes_count']
|
||||
print(f' 生成时间统计特征完毕，特征维度:{time_stat_fea_df.shape}')
|
||||
return time_stat_fea_df
|
||||
|
||||
|
||||
def get_time_std_fea(train, test):
|
||||
print('生成 server_model 特征')
|
||||
df = pd.concat([train, test], axis=0, ignore_index=True)
|
||||
# df['year'] = df['time'].dt.year
|
||||
# df['month'] = df['time'].dt.month
|
||||
df['hour'] = df['time'].dt.hour
|
||||
# df['week'] = df['time'].dt.week
|
||||
df['minute'] = df['time'].dt.minute
|
||||
time_std = df.groupby(['sn', 'server_model']).agg(
|
||||
{'hour': 'std', 'minute': 'std'}).reset_index()
|
||||
time_std = time_std.rename(
|
||||
columns={
|
||||
'hour': 'hour_std',
|
||||
'minute': 'minute_std'})
|
||||
return time_std
|
||||
|
||||
|
||||
def get_key(all_data):
|
||||
all_data['msg_list'] = all_data['msg'].apply(lambda x: [i.strip() for i in x.split(' | ')])
|
||||
class_fea_cnt_list = []
|
||||
for label in [0,1,2,3]:
|
||||
class_df = all_data.query(f'label =={label}')
|
||||
counter = Counter()
|
||||
for i in class_df['msg_list']:
|
||||
counter.update(i)
|
||||
class_fea_cnt = pd.DataFrame({i[0]:i[1] for i in counter.most_common()},index = [f'fea_cnt_{label}']).T.reset_index().rename(columns = {'index':'fea'})
|
||||
class_fea_cnt_list.append(class_fea_cnt)
|
||||
|
||||
fea_cnt_df = class_fea_cnt_list[0]
|
||||
for tmp in class_fea_cnt_list[1:]:
|
||||
fea_cnt_df = fea_cnt_df.merge(tmp,on = 'fea')
|
||||
|
||||
fea_cnt_df['fea_cnt_sum'] = fea_cnt_df.loc[:,['fea_cnt_0', 'fea_cnt_1', 'fea_cnt_2', 'fea_cnt_3']].sum(1)
|
||||
|
||||
all_fea_cnt = fea_cnt_df['fea_cnt_sum'].sum()
|
||||
|
||||
for i in ['fea_cnt_0', 'fea_cnt_1', 'fea_cnt_2', 'fea_cnt_3']:
|
||||
fea_cnt_df[f'{i}_ratio'] = fea_cnt_df[i]/fea_cnt_df['fea_cnt_sum']
|
||||
fea_cnt_df[f'{i}_all_ratio'] = fea_cnt_df[i]/all_fea_cnt
|
||||
|
||||
fea_cnt_df['fea_cnt_ratio_std'] = fea_cnt_df.loc[:,['fea_cnt_0_ratio','fea_cnt_1_ratio','fea_cnt_2_ratio','fea_cnt_3_ratio', ]].std(1)
|
||||
fea_cnt_df['fea_cnt_std'] = fea_cnt_df.loc[:,['fea_cnt_0', 'fea_cnt_1','fea_cnt_2','fea_cnt_3',]].std(1)
|
||||
|
||||
fea_cnt_df['fea_cnt_all_ratio_std'] = fea_cnt_df.loc[:,['fea_cnt_0_all_ratio','fea_cnt_1_all_ratio',
|
||||
'fea_cnt_2_all_ratio','fea_cnt_3_all_ratio',]].std(1)
|
||||
|
||||
fea_cnt_df = fea_cnt_df[~fea_cnt_df['fea_cnt_ratio_std'].isnull()].sort_values('fea_cnt_ratio_std',ascending = False)
|
||||
|
||||
fea_cnt_df['fea_max'] = np.argmax(fea_cnt_df.loc[:,['fea_cnt_0', 'fea_cnt_1', 'fea_cnt_2', 'fea_cnt_3',]].values,axis = 1)
|
||||
key_0 = fea_cnt_df.query('fea_max ==0 ')['fea'].to_list()
|
||||
key_1 = fea_cnt_df.query('fea_max ==1 ')['fea'].to_list()
|
||||
key_2 = fea_cnt_df.query('fea_max ==2 ')['fea'].to_list()
|
||||
key_3 = fea_cnt_df.query('fea_max ==3 ')['fea'].to_list()
|
||||
# key_1 = ['OEM record c2','Processor CPU_Core_Error','001c4c','System Event Sys_Event','Power Supply PS0_Status','Temperature CPU0_Margin_Temp','Reading 51 > Threshold 85 degrees C','Lower Non-critical going low','Temperature CPU1_Margin_Temp','System ACPI Power State #0x7d','Lower Critical going low']
|
||||
# key_2 = ['OEM CPU0 MCERR','OEM CPU0 CATERR','Reading 0 < Threshold 2 degrees C','0203c0a80101','Unknown CPU0 MCERR','Unknown CPU0 CATERR','Microcontroller #0x3b','System Boot Initiated','Processor #0xfa','Power Unit Pwr Unit Status','Hard reset','Power off/down','System Event #0xff','Memory CPU1A1_DIMM_Stat','000000','Power cycle','OEM record c3','Memory CPU1C0_DIMM_Stat','Reading 0 < Threshold 1 degrees C','IERR']
|
||||
# key_3 = ['Memory','Correctable ECC logging limit reached','Memory MEM_CHE0_Status','Memory Memory_Status','Memory #0x87','Memory CPU0F0_DIMM_Stat','Memory Device Disabled','Memory #0xe2','OS Stop/Shutdown OS Status','System Boot Initiated System Restart','OS Boot BIOS_Boot_Up','System Boot Initiated BIOS_Boot_UP','Memory DIMM101','OS graceful shutdown','OS Critical Stop OS Status','Memory #0xf9','Memory CPU0C0_DIMM_Stat','Memory DIMM111','Memory DIMM021',]
|
||||
# key_4 = ['Drive Fault','NMI/Diag Interrupt','Failure detected','Power Supply AC lost','Power Supply PSU0_Supply','AC out-of-range, but present','Predictive failure','Drive Present','Temperature Temp_DIMM_KLM','Temperature Temp_DIMM_DEF','Power Supply PS1_Status','Identify Status','Power Supply PS2_Status','Temperature DIMMG1_Temp','Upper Non-critical going high','Temperature DIMMG0_Temp','Upper Critical going high','Power Button pressed','System Boot Initiated #0xb8','Deasserted']
|
||||
return key_0,key_1,key_2,key_3
|
||||
|
||||
def get_class_key_words_nunique(all_data):
|
||||
print('获取 class_key_words_nunique 特征')
|
||||
|
||||
key_0,key_1,key_2,key_3 = get_key(all_data)
|
||||
|
||||
df = all_data[['sn', 'fault_time', 'msg_list']]
|
||||
df_tmp = df.groupby(['sn' ]).agg({'msg_list':'sum'}).reset_index()
|
||||
df_tmp['class_0_key_words_nunique'] = df_tmp['msg_list'].apply(lambda x:len(set(x)&set(key_0)))
|
||||
df_tmp['class_1_key_words_nunique'] = df_tmp['msg_list'].apply(lambda x:len(set(x)&set(key_1)))
|
||||
df_tmp['class_2_key_words_nunique'] = df_tmp['msg_list'].apply(lambda x:len(set(x)&set(key_2)))
|
||||
df_tmp['class_3_key_words_nunique'] = df_tmp['msg_list'].apply(lambda x:len(set(x)&set(key_3)))
|
||||
del df_tmp['msg_list']
|
||||
return df_tmp
|
||||
def get_key_for_top_fea(train,test):
|
||||
KEY_FOR_TOP_COLS = []
|
||||
print('添加 key_for_top_fea 特征')
|
||||
for TIME in [3, 5, 10, 15, 30, 45, 60, 90, 120, 240, 360, 480, 540, 600,60000000]:
|
||||
for i in range(10):
|
||||
train[f'KEY_FOR_TOP_{i}_{TIME}'] = train[f'{KEY_1[i]}_{TIME}'].astype(str)+'_'+train[f'{KEY_2[i]}_{TIME}'].astype(str)+'_'+train[f'{KEY_3[i]}_{TIME}'].astype(str)+'_'+train[f'{KEY_4[i]}_{TIME}'].astype(str)
|
||||
test[f'KEY_FOR_TOP_{i}_{TIME}'] = test[f'{KEY_1[i]}_{TIME}'].astype(str)+'_'+test[f'{KEY_2[i]}_{TIME}'].astype(str)+'_'+test[f'{KEY_3[i]}_{TIME}'].astype(str)+'_'+test[f'{KEY_4[i]}_{TIME}'].astype(str)
|
||||
KEY_FOR_TOP_COLS.append(f'KEY_FOR_TOP_{i}_{TIME}')
|
||||
train = cat2num(train,KEY_FOR_TOP_COLS)
|
||||
test = cat2num(test,KEY_FOR_TOP_COLS)
|
||||
for KEY_FOR_TOP_COL in KEY_FOR_TOP_COLS:
|
||||
del train[KEY_FOR_TOP_COL]
|
||||
del test[KEY_FOR_TOP_COL]
|
||||
return train,test
|
||||
|
||||
def get_key_word_cross_fea(train,test):
|
||||
print('获取关键词交叉特征......')
|
||||
KEY_WORDS_MAP = {'CPU0':KEY_1,'CPU1':KEY_2,'CPU2':KEY_3,'CPU3':KEY_4}
|
||||
KEY_WORDS_CROSS_COLS =[]
|
||||
for KEY_WORDS in KEY_WORDS_MAP:
|
||||
for i in [3, 5, 10, 15, 30, 45, 60, 90, 120, 240, 360, 480, 540, 600,60000000]:
|
||||
KEY_WORDS_COLS = [f'{col}_{i}' for col in KEY_WORDS_MAP[KEY_WORDS]]
|
||||
train[f'{KEY_WORDS}_WORDS_{i}'] = train[KEY_WORDS_COLS].astype(str).sum(1)
|
||||
test[f'{KEY_WORDS}_WORDS_{i}'] = test[KEY_WORDS_COLS].astype(str).sum(1)
|
||||
KEY_WORDS_CROSS_COLS.append(f'{KEY_WORDS}_WORDS_{i}')
|
||||
train = cat2num(train,KEY_WORDS_CROSS_COLS)
|
||||
test = cat2num(test,KEY_WORDS_CROSS_COLS)
|
||||
|
||||
for COLS in KEY_WORDS_CROSS_COLS:
|
||||
del train[COLS]
|
||||
del test[COLS]
|
||||
print('获取关键词交叉特征完毕......')
|
||||
return train,test
|
||||
def get_time_quantile_fea(df):
|
||||
print(' 生成时间分位数特征')
|
||||
secs = [0.2, 0.4, 0.6, 0.8]
|
||||
time_fea_list = []
|
||||
for sec in tqdm(secs):
|
||||
for time_fea_type in [
|
||||
'duration_minutes', 'log_duration_minutes', 'time_diff_1', 'log_time_diff_1']:
|
||||
temp = df.groupby(['sn', 'server_model', 'fault_time'])[time_fea_type].quantile(sec).reset_index(
|
||||
name=f'{time_fea_type}_' + str(sec * 100))
|
||||
|
||||
time_fea_list.append(temp)
|
||||
time_fea_df = time_fea_list[0]
|
||||
for time_fea in time_fea_list[1:]:
|
||||
time_fea_df = time_fea_df.merge(
|
||||
time_fea, how='left', on=[
|
||||
'sn', 'server_model', 'fault_time'])
|
||||
print(f' 生成时间分位数特征完毕，特征维度:{time_fea_df.shape}')
|
||||
return time_fea_df
|
||||
|
||||
|
||||
def get_server_model_fea(train, test):
|
||||
print('生成 server_model 特征')
|
||||
df = pd.concat([train, test], axis=0, ignore_index=True)
|
||||
df['server_model_count_sn'] = df.groupby(
|
||||
['server_model'])['sn'].transform('count')
|
||||
df['server_model_nunique_sn'] = df.groupby(
|
||||
['server_model'])['sn'].transform('nunique')
|
||||
# df['server_model_count'] = df.groupby('server_model')['server_model'].transform('count')
|
||||
# df['server_model_cnt_quantile'] = df['server_model'].map(
|
||||
# df['server_model'].value_counts().rank() / len(df['server_model'].unique()))
|
||||
# df['server_model_cnt_rank'] = df[f'server_model_cnt_quantile'].rank(method='min')
|
||||
|
||||
df['sn_cnt'] = df['sn'].map(df['sn'].value_counts())
|
||||
df['sn_freq'] = df['sn'].map(df['sn'].value_counts() / len(df))
|
||||
df['server_model_cnt'] = df['server_model'].map(
|
||||
df['server_model'].value_counts())
|
||||
df['server_model_freq'] = df['server_model'].map(
|
||||
df['server_model'].value_counts() / len(df))
|
||||
select_cols = ['sn', 'server_model',
|
||||
'server_model_count_sn', 'server_model_nunique_sn',
|
||||
'sn_cnt', 'sn_freq', 'server_model_cnt', 'server_model_freq'
|
||||
# 'server_model_count','server_model_cnt_quantile', 'server_model_cnt_rank'
|
||||
]
|
||||
server_model_fea = df[select_cols]
|
||||
|
||||
cat_feats = [
|
||||
'server_model'] # ,'server_model_day_date','server_model_dayofmonth','server_model_dayofweek','server_model_hour']
|
||||
# for name in cat_feats:
|
||||
# le = LabelEncoder()
|
||||
# server_model_fea[f'{name}_LabelEnc'] = le.fit_transform(
|
||||
# server_model_fea[name])
|
||||
server_model_fea = cat2num(server_model_fea, cat_feats, Transfer2num=True)
|
||||
server_model_fea = server_model_fea.drop_duplicates().reset_index(drop=True)
|
||||
print(f'生成 server_model 特征完毕，特征维度:{server_model_fea.shape}')
|
||||
|
||||
return server_model_fea
|
||||
|
||||
|
||||
def get_time_type_msg_unique_fea(df):
|
||||
df['msg_list'] = df['msg'].apply(
|
||||
lambda x: [i.strip() for i in x.split(' | ')])
|
||||
|
||||
df['msg_0'] = df['msg'].apply(
|
||||
lambda x: [
|
||||
get_msg_location(
|
||||
x.split(' | '),
|
||||
0)])
|
||||
df['msg_1'] = df['msg'].apply(
|
||||
lambda x: [
|
||||
get_msg_location(
|
||||
x.split(' | '),
|
||||
1)])
|
||||
df['msg_2'] = df['msg'].apply(
|
||||
lambda x: [
|
||||
get_msg_location(
|
||||
x.split(' | '),
|
||||
2)])
|
||||
|
||||
df = df.groupby(['sn', 'fault_time']).agg(
|
||||
{'msg_list': 'sum', 'msg_0': 'sum', 'msg_1': 'sum', 'msg_2': 'sum'}).reset_index()
|
||||
|
||||
df['msg_set'] = df['msg_list'].apply(lambda x: '|'.join(list(set(x))))
|
||||
|
||||
df['msg_0_set'] = df['msg_0'].apply(lambda x: '|'.join(list(set(x))))
|
||||
df['msg_1_set'] = df['msg_1'].apply(lambda x: '|'.join(list(set(x))))
|
||||
df['msg_2_set'] = df['msg_2'].apply(lambda x: '|'.join(list(set(x))))
|
||||
df = df[['sn', 'fault_time', 'msg_set',
|
||||
'msg_0_set', 'msg_1_set', 'msg_2_set']]
|
||||
return df
|
||||
|
||||
|
||||
def get_msg_unique_fea(train, test, time_type='last'):
|
||||
print('生成msg_unique_ fea')
|
||||
common_cols = ['msg_set', 'msg_0_set', 'msg_1_set', 'msg_2_set']
|
||||
df = pd.concat([train, test], axis=0, ignore_index=True)
|
||||
df['time_interval'] = (
|
||||
pd.to_datetime(
|
||||
df['fault_time']) -
|
||||
df['time']).apply(
|
||||
lambda x: x.total_seconds())
|
||||
|
||||
last_fea = get_time_type_msg_unique_fea(df.query('time_interval >0'))
|
||||
last_fea = last_fea.rename(columns={i: f'last_{i}' for i in common_cols})
|
||||
next_fea = get_time_type_msg_unique_fea(df.query('time_interval <0'))
|
||||
next_fea = next_fea.rename(columns={i: f'next_{i}' for i in common_cols})
|
||||
all_fea = get_time_type_msg_unique_fea(df)
|
||||
all_fea = all_fea.rename(columns={i: f'all_{i}' for i in common_cols})
|
||||
msg_unique_fea = all_fea.merge(
|
||||
last_fea, on=['sn', 'fault_time'], how='outer')
|
||||
msg_unique_fea = msg_unique_fea.merge(
|
||||
next_fea, on=['sn', 'fault_time'], how='outer')
|
||||
return msg_unique_fea
|
||||
|
||||
|
||||
def get_duration_minutes_fea(train, test):
|
||||
print('生成 duration_minutes 特征')
|
||||
df = pd.concat([train, test], axis=0, ignore_index=True)
|
||||
df['duration_minutes'] = (pd.to_datetime(df['fault_time']) - pd.to_datetime(df['time'])).apply(
|
||||
lambda x: x.total_seconds())
|
||||
df['log_duration_minutes'] = np.log(df['duration_minutes'])
|
||||
|
||||
df = df.sort_values(['sn', 'label', 'server_model',
|
||||
'fault_time', 'time']).reset_index(drop=True)
|
||||
df['time_diff_1'] = (df.groupby(['sn', 'server_model', 'fault_time'])['time'].diff(1)).apply(
|
||||
lambda x: x.total_seconds())
|
||||
df['time_diff_1'] = df['time_diff_1'].fillna(0)
|
||||
df['log_time_diff_1'] = np.log(df['time_diff_1'])
|
||||
|
||||
# time_quantile_fea_df = get_time_quantile_fea(df)
|
||||
# time_stat_fea_df = get_4_time_stat_fea(df)
|
||||
# df_tmp = time_quantile_fea_df.merge(time_stat_fea_df, on= ['sn', 'server_model','fault_time'],how = 'left')
|
||||
time_stat_fea_df = get_4_time_stat_fea(df)
|
||||
df_tmp = time_stat_fea_df
|
||||
print(f'生成 duration_minutes 特征完毕，特征维度{df_tmp.shape}')
|
||||
return df_tmp
|
||||
|
||||
|
||||
def get_msg_text_fea_all(all_data):
|
||||
all_data['label'] = all_data['label'].fillna(-1)
|
||||
all_data['msg_list'] = all_data['msg'].apply(lambda x: [i.strip() for i in x.split(' | ')])
|
||||
all_data['msg_0'] = all_data['msg'].apply(lambda x: [get_msg_location(x.split(' | '), 0)])
|
||||
all_data['msg_1'] = all_data['msg'].apply(lambda x: [get_msg_location(x.split(' | '), 1)])
|
||||
all_data['msg_2'] = all_data['msg'].apply(lambda x: [get_msg_location(x.split(' | '), 2)])
|
||||
|
||||
all_data = all_data.sort_values(['sn', 'fault_time', 'time']).reset_index(drop=True)
|
||||
del all_data['label']
|
||||
last_data = all_data.query('time_interval >0')
|
||||
next_data = all_data.query('time_interval <=0')
|
||||
|
||||
# id_cols = ['sn', 'fault_time', 'label']
|
||||
|
||||
# all_msg_text_fea = get_msg_text_fea(all_data, msg_type='all')
|
||||
last_msg_text_fea = get_msg_text_fea(last_data, msg_type='last')
|
||||
# next_msg_text_fea = get_msg_text_fea(next_data, msg_type='next')
|
||||
msg_text_fea = last_msg_text_fea
|
||||
return msg_text_fea
|
||||
|
||||
def get_test_key_words(train,test):
|
||||
|
||||
df = pd.concat([train[['sn', 'fault_time', 'label','msg']],test[['sn', 'fault_time', 'msg']]],ignore_index = True).drop_duplicates(['sn', 'fault_time', 'msg'])
|
||||
df['label'] = df['label'].fillna(5)
|
||||
df['msg_list'] = df['msg'].apply(lambda x:[i.strip() for i in x.split(' | ')])
|
||||
words_cnt_df_list = []
|
||||
for label in df['label'].unique():
|
||||
label = int(label)
|
||||
df_tmp = df.query(f'label == {label}')
|
||||
counter = Counter()
|
||||
for words in df_tmp['msg_list']:
|
||||
words = [i.replace('_',' ') for i in words]
|
||||
# word_list = []
|
||||
# for i in words:
|
||||
# word_list+=i.split(' ')
|
||||
# words = word_list
|
||||
counter.update(words)
|
||||
words_cnt_df = pd.DataFrame(counter,index = [0]).T.reset_index().rename(columns = {'index':'word',0:f'cnt_{label}'})
|
||||
words_cnt_df_list.append(words_cnt_df)
|
||||
words_cnt_df = words_cnt_df_list[0]
|
||||
for i in words_cnt_df_list[1:]:
|
||||
words_cnt_df = words_cnt_df.merge(i,on = 'word',how = 'outer' )
|
||||
|
||||
words_cnt_df = words_cnt_df.fillna(-1)
|
||||
words_cnt_df1 = words_cnt_df.query('cnt_0 >10 and cnt_2 >10 and cnt_1 >10 and cnt_3>10 and cnt_5>10 ')
|
||||
cnt_class = ['cnt_0','cnt_1','cnt_2','cnt_3','cnt_5']
|
||||
words_cnt_df1['word_cnt_sum'] = words_cnt_df1.loc[:,cnt_class].sum(1)
|
||||
for i in cnt_class:
|
||||
words_cnt_df1[f'{i}_ratio'] = words_cnt_df1[i]/words_cnt_df1['word_cnt_sum']
|
||||
words_cnt_df1['word_cnt_ratio_std'] = words_cnt_df1.loc[:,['cnt_0_ratio','cnt_1_ratio', 'cnt_2_ratio', 'cnt_3_ratio']].std(1)
|
||||
words_cnt_df1['cnt_1_0_diff'] = (words_cnt_df1['cnt_1_ratio'] - words_cnt_df1['cnt_0_ratio'])
|
||||
test_key_words = words_cnt_df1.sort_values('cnt_5',ascending = False)['word'].to_list()[5:40]
|
||||
return test_key_words
|
||||
|
||||
def get_w2v_mean(w2v_model,sentences):
|
||||
emb_matrix = list()
|
||||
vec = list()
|
||||
for w in sentences.split():
|
||||
if w in w2v_model.wv:
|
||||
vec.append(w2v_model.wv[w])
|
||||
if len(vec) > 0:
|
||||
emb_matrix.append(np.mean(vec, axis=0))
|
||||
else:
|
||||
emb_matrix.append([0] * w2v_model.vector_size)
|
||||
return emb_matrix
|
||||
def get_tfidf_svd(tfv,svd,sentences, n_components=16):
|
||||
X_tfidf = tfv.transform(sentences)
|
||||
X_svd = svd.transform(X_tfidf)
|
||||
return np.mean(X_svd, axis=0)
|
||||
def get_w2v_tfidf_fea(all_data):
|
||||
print('w2v编码')
|
||||
df = all_data
|
||||
df['msg_list'] = df['msg'].apply(lambda x: [i.strip().lower().replace(' ','_') for i in x.split(" | ")])
|
||||
df = df.groupby(['sn']).agg({'msg_list': 'sum'}).reset_index()
|
||||
df['text'] = df['msg_list'].apply(lambda x: ' '.join(x))
|
||||
|
||||
sentences_list = df['text'].values.tolist()
|
||||
sentences = []
|
||||
for s in sentences_list:
|
||||
sentences.append([w for w in s.split()])
|
||||
w2v_model = Word2Vec(sentences, vector_size=10, window=3, min_count=5, sg=0, hs=1, seed=2022)
|
||||
df['text_w2v'] = df['text'].apply(lambda x: get_w2v_mean(w2v_model, x)[0])
|
||||
|
||||
print('tfidf编码')
|
||||
X = df['text'].to_list()
|
||||
tfv = TfidfVectorizer(ngram_range=(1, 3), min_df=5, max_features=50000)
|
||||
tfv.fit(X)
|
||||
X_tfidf = tfv.transform(X)
|
||||
svd = TruncatedSVD(n_components=16) # 降维
|
||||
svd.fit(X_tfidf)
|
||||
df['text_tfidf'] = df['text'].apply(lambda x: get_tfidf_svd(tfv, svd, x.split()))
|
||||
|
||||
print("doc2vec编码")
|
||||
texts = df['text'].tolist()
|
||||
documents = [TaggedDocument(doc, [i]) for i, doc in enumerate(texts)]
|
||||
model = Doc2Vec(documents, window=5, min_count=3, workers=4)
|
||||
docvecs = model.docvecs
|
||||
df['doc2vec'] = [docvecs[i] for i in range(len(docvecs))]
|
||||
|
||||
for i in range(32):
|
||||
df[f'msg_w2v_{i}'] = df['text_w2v'].apply(lambda x: x[i])
|
||||
for i in range(16):
|
||||
df[f'msg_tfv_{i}'] = df['text_tfidf'].apply(lambda x: x[i])
|
||||
for i in range(100):
|
||||
df[f'msg_doc2vec_{i}'] = df['doc2vec'].apply(lambda x: x[i])
|
||||
|
||||
save_cols = [i for i in df.columns if i not in ['msg_list', 'text', 'text_w2v', 'text_tfidf', 'doc2vec']]
|
||||
return df[save_cols]
|
||||
|
||||
# w2v_tfidf_fea = get_w2v_tfidf_fea(all_data)
|
||||
class BetaEncoder(object):
|
||||
|
||||
def __init__(self, group):
|
||||
|
||||
self.group = group
|
||||
self.stats = None
|
||||
|
||||
# get counts from df
|
||||
def fit(self, df, target_col):
|
||||
# 先验均值
|
||||
self.prior_mean = np.mean(df[target_col])
|
||||
stats = df[[target_col, self.group]].groupby(self.group)
|
||||
# count和sum
|
||||
stats = stats.agg(['sum', 'count'])[target_col]
|
||||
stats.rename(columns={'sum': 'n', 'count': 'N'}, inplace=True)
|
||||
stats.reset_index(level=0, inplace=True)
|
||||
self.stats = stats
|
||||
|
||||
# extract posterior statistics
|
||||
def transform(self, df, stat_type, N_min=1):
|
||||
|
||||
df_stats = pd.merge(df[[self.group]], self.stats, how='left')
|
||||
n = df_stats['n'].copy()
|
||||
N = df_stats['N'].copy()
|
||||
|
||||
# fill in missing
|
||||
nan_indexs = np.isnan(n)
|
||||
n[nan_indexs] = self.prior_mean
|
||||
N[nan_indexs] = 1.0
|
||||
|
||||
# prior parameters
|
||||
N_prior = np.maximum(N_min - N, 0)
|
||||
alpha_prior = self.prior_mean * N_prior
|
||||
beta_prior = (1 - self.prior_mean) * N_prior
|
||||
|
||||
# posterior parameters
|
||||
alpha = alpha_prior + n
|
||||
beta = beta_prior + N - n
|
||||
|
||||
# calculate statistics
|
||||
if stat_type == 'mean':
|
||||
num = alpha
|
||||
dem = alpha + beta
|
||||
|
||||
elif stat_type == 'mode':
|
||||
num = alpha - 1
|
||||
dem = alpha + beta - 2
|
||||
|
||||
elif stat_type == 'median':
|
||||
num = alpha - 1 / 3
|
||||
dem = alpha + beta - 2 / 3
|
||||
|
||||
elif stat_type == 'var':
|
||||
num = alpha * beta
|
||||
dem = (alpha + beta) ** 2 * (alpha + beta + 1)
|
||||
|
||||
elif stat_type == 'skewness':
|
||||
num = 2 * (beta - alpha) * np.sqrt(alpha + beta + 1)
|
||||
dem = (alpha + beta + 2) * np.sqrt(alpha * beta)
|
||||
|
||||
elif stat_type == 'kurtosis':
|
||||
num = 6 * (alpha - beta) ** 2 * (alpha + beta + 1) - \
|
||||
alpha * beta * (alpha + beta + 2)
|
||||
dem = alpha * beta * (alpha + beta + 2) * (alpha + beta + 3)
|
||||
|
||||
# replace missing
|
||||
value = num / dem
|
||||
value[np.isnan(value)] = np.nanmedian(value)
|
||||
return value
|
||||
|
||||
|
||||
def get_beta_target(train, test):
|
||||
N_min = 1000
|
||||
feature_cols = []
|
||||
|
||||
# encode variables
|
||||
for c in ['server_model']:
|
||||
# fit encoder
|
||||
be = BetaEncoder(c)
|
||||
be.fit(train, 'label')
|
||||
|
||||
# mean
|
||||
feature_name = f'{c}_mean'
|
||||
train[feature_name] = be.transform(train, 'mean', N_min)
|
||||
test[feature_name] = be.transform(test, 'mean', N_min)
|
||||
feature_cols.append(feature_name)
|
||||
|
||||
# mode
|
||||
feature_name = f'{c}_mode'
|
||||
train[feature_name] = be.transform(train, 'mode', N_min)
|
||||
test[feature_name] = be.transform(test, 'mode', N_min)
|
||||
feature_cols.append(feature_name)
|
||||
|
||||
# median
|
||||
feature_name = f'{c}_median'
|
||||
train[feature_name] = be.transform(train, 'median', N_min)
|
||||
test[feature_name] = be.transform(test, 'median', N_min)
|
||||
feature_cols.append(feature_name)
|
||||
|
||||
# var
|
||||
feature_name = f'{c}_var'
|
||||
train[feature_name] = be.transform(train, 'var', N_min)
|
||||
test[feature_name] = be.transform(test, 'var', N_min)
|
||||
feature_cols.append(feature_name)
|
||||
|
||||
# # skewness
|
||||
# feature_name = f'{c}_skewness'
|
||||
# train[feature_name] = be.transform(train, 'skewness', N_min)
|
||||
# test[feature_name] = be.transform(test, 'skewness', N_min)
|
||||
# feature_cols.append(feature_name)
|
||||
|
||||
# kurtosis
|
||||
feature_name = f'{c}_kurtosis'
|
||||
train[feature_name] = be.transform(train, 'kurtosis', N_min)
|
||||
test[feature_name] = be.transform(test, 'kurtosis', N_min)
|
||||
feature_cols.append(feature_name)
|
||||
df = train.append(test).reset_index(drop=True)
|
||||
df = df[['sn', 'fault_time', 'server_model', 'server_model_mean',
|
||||
'server_model_mode', 'server_model_median', 'server_model_var',
|
||||
'server_model_kurtosis']].drop_duplicates().reset_index(drop=True)
|
||||
return df
|
||||
@ -0,0 +1,73 @@
|
||||
import pandas as pd
|
||||
import numpy as np
|
||||
import os
|
||||
from utils import TRAIN_DIR ,TEST_A_DIR,TEST_B_DIR,RESULT_DIR,DATA_DIR
|
||||
|
||||
log_dataset_a = pd.read_csv(os.path.join(DATA_DIR,'preliminary_a_test/preliminary_sel_log_dataset_a.csv'))
|
||||
log_dataset_b = pd.read_csv(os.path.join(DATA_DIR,'preliminary_b_test/preliminary_sel_log_dataset_b.csv'))
|
||||
submit_dataset_a = pd.read_csv(os.path.join(DATA_DIR,'preliminary_a_test/preliminary_submit_dataset_a.csv'))
|
||||
submit_dataset_b = pd.read_csv(os.path.join(DATA_DIR,'preliminary_b_test/preliminary_submit_dataset_b.csv'))
|
||||
|
||||
log_dataset_c = pd.concat([log_dataset_a,log_dataset_b],ignore_index = True,axis = 0)
|
||||
submit_dataset_c = pd.concat([submit_dataset_a,submit_dataset_b],ignore_index = True,axis = 0)
|
||||
|
||||
log_dataset_c.to_csv(os.path.join(TEST_A_DIR,'final_log_dataset_c.csv'),index =False)
|
||||
submit_dataset_c.to_csv(os.path.join(TEST_A_DIR,'final_submit_dataset_c.csv'),index =False)
|
||||
|
||||
|
||||
#
|
||||
# cat_prob = pd.read_csv(os.path.join(RESULT_DIR,'../../../TianchiAIOps_bert_model/cat_prob_result.csv'))
|
||||
# lgb_prob = pd.read_csv(os.path.join(RESULT_DIR,'../../../TianchiAIOps_bert_model/lgb_prob_result.csv'))
|
||||
|
||||
cat_prob = pd.read_csv(os.path.join(RESULT_DIR,'B_prob_7511.csv'))
|
||||
lgb_prob = pd.read_csv(os.path.join(RESULT_DIR,'baseline_prob_7495.csv'))
|
||||
cat_prob.columns = ['cat_class_0', 'cat_class_1', 'cat_class_2', 'cat_class_3', 'label', 'sn',
|
||||
'fault_time']
|
||||
lgb_prob.columns = ['lgb_class_0', 'lgb_class_1', 'lgb_class_2', 'lgb_class_3', 'label', 'sn',
|
||||
'fault_time']
|
||||
|
||||
lgb_prob = lgb_prob[lgb_prob['label'].isnull()]
|
||||
cat_prob = cat_prob[cat_prob['label'].isnull()]
|
||||
|
||||
cat_prob['cat_prob'] = cat_prob.loc[:,['cat_class_0', 'cat_class_1', 'cat_class_2', 'cat_class_3']].max(1)
|
||||
cat_prob['cat_label'] = np.argmax(cat_prob.loc[:,['cat_class_0', 'cat_class_1', 'cat_class_2', 'cat_class_3']].values,axis = 1)
|
||||
|
||||
lgb_prob['lgb_prob'] = lgb_prob.loc[:,['lgb_class_0', 'lgb_class_1', 'lgb_class_2', 'lgb_class_3']].max(1)
|
||||
lgb_prob['lgb_label'] = np.argmax(lgb_prob.loc[:,['lgb_class_0', 'lgb_class_1', 'lgb_class_2', 'lgb_class_3']].values,axis = 1)
|
||||
|
||||
lgb_prob = lgb_prob[['sn','fault_time','lgb_label','lgb_prob']]
|
||||
cat_prob = cat_prob[['sn','fault_time','cat_label','cat_prob']]
|
||||
|
||||
# prob = cat_prob.merge(lgb_prob,on =['sn','fault_time'],
|
||||
# how = 'left' )
|
||||
|
||||
prob = pd.concat([cat_prob,lgb_prob],ignore_index = True)
|
||||
prob['cat_prob']=prob['cat_prob'].fillna(1)
|
||||
prob['lgb_prob']=prob['lgb_prob'].fillna(1)
|
||||
prob.loc[prob['cat_label'].isnull(),'cat_label'] = prob.loc[prob['cat_label'].isnull(),'lgb_label']
|
||||
prob.loc[prob['lgb_label'].isnull(),'lgb_label'] = prob.loc[prob['lgb_label'].isnull(),'cat_label']
|
||||
|
||||
|
||||
pseudo_labels = prob.query('cat_prob >0.85 and lgb_prob >0.85 and lgb_label == cat_label ')
|
||||
|
||||
pseudo_labels = pseudo_labels[['sn','fault_time','cat_label']].rename(columns = {'cat_label':'label'}).reset_index(drop = True)
|
||||
pseudo_labels.to_csv(os.path.join(TRAIN_DIR,'pseudo_labels.csv'),index= False)
|
||||
print(f'生成伪标签的数据维度:{pseudo_labels.shape}')
|
||||
|
||||
pseudo_sel_log_dataset = pd.read_csv(os.path.join(TEST_A_DIR,'final_sel_log_dataset_c.csv'))
|
||||
pseudo_sel_log_dataset = pseudo_sel_log_dataset[pseudo_sel_log_dataset['sn'].isin(pseudo_labels['sn'].to_list())]
|
||||
pseudo_sel_log_dataset.to_csv(os.path.join(TRAIN_DIR,'pseudo_sel_log_dataset.csv'),index = False)
|
||||
print(f'生成伪标签的日志数据维度:{pseudo_sel_log_dataset.shape}')
|
||||
|
||||
# 制作新的测试集
|
||||
final_submit_dataset_d= prob.merge(pseudo_labels,on =['sn','fault_time'],how = 'left' )
|
||||
final_submit_dataset_d = final_submit_dataset_d[final_submit_dataset_d['label'].isnull()][['sn','fault_time' ]].reset_index(drop = True)
|
||||
final_submit_dataset_d.to_csv(os.path.join(TEST_A_DIR,'final_submit_dataset_d.csv'),index= False)
|
||||
print(f'生成新的测试集维度:{final_submit_dataset_d.shape}')
|
||||
|
||||
final_sel_log_dataset_d = pd.read_csv(os.path.join(TEST_A_DIR,'final_sel_log_dataset_c.csv'))
|
||||
final_sel_log_dataset_d = final_sel_log_dataset_d[final_sel_log_dataset_d['sn'].isin(final_submit_dataset_d['sn'].to_list())]
|
||||
|
||||
final_sel_log_dataset_d.to_csv(
|
||||
os.path.join(TEST_A_DIR,'final_sel_log_dataset_d.csv'),index = False)
|
||||
print(f'生成新的测试集日志数据维度:{final_sel_log_dataset_d.shape}')
|
||||
@ -0,0 +1,291 @@
|
||||
import os
|
||||
import warnings
|
||||
|
||||
import numpy as np
|
||||
import pandas as pd
|
||||
import datetime
|
||||
from generate_feature import get_beta_target, add_last_next_time4fault, get_feature, \
|
||||
get_duration_minutes_fea, get_nearest_msg_fea, get_server_model_sn_fea_2, \
|
||||
get_server_model_fea, get_msg_text_fea_all, get_key_word_cross_fea, get_server_model_time_interval_stat_fea, \
|
||||
get_w2v_feats
|
||||
from model import run_cbt,run_lgb
|
||||
from utils import RESULT_DIR, TRAIN_DIR, \
|
||||
TEST_A_DIR, KEY_WORDS, get_word_counter, search_weight, macro_f1, TIME_INTERVAL,PSEUDO_FALG,GENERATION_DIR
|
||||
|
||||
warnings.filterwarnings('ignore')
|
||||
|
||||
|
||||
def get_label(PSEUDO_FALG):
|
||||
preliminary_train_label_dataset = pd.read_csv(preliminary_train_label_dataset_path)
|
||||
preliminary_train_label_dataset_s = pd.read_csv(preliminary_train_label_dataset_s_path)
|
||||
|
||||
if PSEUDO_FALG:
|
||||
print('获取伪标签LABEL')
|
||||
pseudo_labels = pd.read_csv(os.path.join(TRAIN_DIR, 'pseudo_labels.csv'))
|
||||
label = pd.concat([preliminary_train_label_dataset,
|
||||
pseudo_labels,
|
||||
preliminary_train_label_dataset_s],
|
||||
ignore_index=True,
|
||||
axis=0).sort_values(
|
||||
['sn', 'fault_time']).reset_index(drop=True)
|
||||
else:
|
||||
print('不使用伪标签数据')
|
||||
label = pd.concat([preliminary_train_label_dataset,
|
||||
preliminary_train_label_dataset_s],
|
||||
ignore_index=True,
|
||||
axis=0).sort_values(
|
||||
['sn', 'fault_time']).reset_index(drop=True)
|
||||
label['fault_time'] = label['fault_time'].apply(lambda x: datetime.datetime.strptime(x, '%Y-%m-%d %H:%M:%S'))
|
||||
label['fault_time'] = label['fault_time'].apply(lambda x: str(x))
|
||||
return label
|
||||
|
||||
|
||||
def get_log_dateset(PSEUDO_FALG):
|
||||
preliminary_sel_log_dataset = pd.read_csv(preliminary_sel_log_dataset_path)
|
||||
preliminary_sel_log_dataset_a = pd.read_csv(preliminary_sel_log_dataset_a_path)
|
||||
if PSEUDO_FALG:
|
||||
print('获取伪标签日志数据')
|
||||
pseudo_sel_log_dataset = pd.read_csv(os.path.join(TRAIN_DIR, 'pseudo_sel_log_dataset.csv'))
|
||||
log_dataset = pd.concat([preliminary_sel_log_dataset,
|
||||
pseudo_sel_log_dataset,
|
||||
preliminary_sel_log_dataset_a],
|
||||
ignore_index=True,
|
||||
axis=0).sort_values(
|
||||
['sn', 'time', 'server_model']).reset_index(drop=True)
|
||||
else:
|
||||
print('不使用伪标签数据')
|
||||
log_dataset = pd.concat([preliminary_sel_log_dataset,
|
||||
preliminary_sel_log_dataset_a],
|
||||
ignore_index=True,
|
||||
axis=0).sort_values(
|
||||
['sn', 'time', 'server_model']).reset_index(drop=True)
|
||||
log_dataset['time'] = log_dataset['time'].apply(lambda x: datetime.datetime.strptime(x, '%Y-%m-%d %H:%M:%S'))
|
||||
|
||||
return log_dataset
|
||||
|
||||
def get_fea_distribute(feature_df, feature_importances, dataset_type, top=30):
|
||||
print('根据特征重要性，获取数据集的分布情况，用于验证训练集和测试集是否分布一致')
|
||||
fea_distribute_list = []
|
||||
for i in feature_importances[:top]['fea'].to_list():
|
||||
fea_distribute_tmp = (feature_df[i].value_counts() / len(feature_df)).reset_index().rename(
|
||||
columns={'index': 'value'})
|
||||
fea_distribute_list.append(fea_distribute_tmp)
|
||||
|
||||
fea_distribute = fea_distribute_list[-1]
|
||||
for i in fea_distribute_list[:-1]:
|
||||
fea_distribute = fea_distribute.merge(i, on='value', how='left')
|
||||
fea_distribute['value'] = fea_distribute['value'].apply(lambda x: f'{dataset_type}_{int(x)}')
|
||||
return fea_distribute
|
||||
|
||||
|
||||
def get_train_test(label, preliminary_submit_dataset_a, log_dataset):
|
||||
print('获取训练集数据与测试集数据')
|
||||
train = label.merge(log_dataset, on='sn', how='left')
|
||||
test = preliminary_submit_dataset_a.merge(log_dataset, on='sn', how='left')
|
||||
# train['time_interval'] = (pd.to_datetime( train['fault_time'])-train['time'] ).apply(lambda x:x.total_seconds())
|
||||
# test['time_interval'] = (pd.to_datetime( test['fault_time'])- test['time'] ).apply(lambda x:x.total_seconds())
|
||||
# train = train.query('time_interval > 0')
|
||||
# test = test.query('time_interval > 0')
|
||||
print(f'训练集维度:{train.shape},测试集维度:{test.shape}')
|
||||
train = train.drop_duplicates().reset_index(drop=True)
|
||||
test = test.drop_duplicates().reset_index(drop=True)
|
||||
train['time'] = pd.to_datetime(train['time'])
|
||||
test['time'] = pd.to_datetime(test['time'])
|
||||
return train, test
|
||||
|
||||
start_time = datetime.datetime.now()
|
||||
|
||||
additional_sel_log_dataset_path = os.path.join(TRAIN_DIR, 'additional_sel_log_dataset.csv')
|
||||
preliminary_train_label_dataset_path = os.path.join(TRAIN_DIR, 'preliminary_train_label_dataset.csv')
|
||||
preliminary_train_label_dataset_s_path = os.path.join(TRAIN_DIR, 'preliminary_train_label_dataset_s.csv')
|
||||
preliminary_sel_log_dataset_path = os.path.join(TRAIN_DIR, 'preliminary_sel_log_dataset.csv')
|
||||
|
||||
preliminary_submit_dataset_a_path = os.path.join(TEST_A_DIR, 'final_submit_dataset_b.csv')
|
||||
preliminary_sel_log_dataset_a_path = os.path.join(TEST_A_DIR, 'final_sel_log_dataset_b.csv')
|
||||
|
||||
print(preliminary_submit_dataset_a_path, preliminary_sel_log_dataset_a_path)
|
||||
|
||||
preliminary_submit_dataset_a = pd.read_csv(preliminary_submit_dataset_a_path)
|
||||
preliminary_submit_dataset_a.head()
|
||||
|
||||
log_dataset = get_log_dateset(PSEUDO_FALG)
|
||||
label = get_label(PSEUDO_FALG)
|
||||
|
||||
|
||||
next_time_list = [i / TIME_INTERVAL for i in [3, 5, 10, 15, 30, 45, 60, 90, 120, 240, 360, 480, 540, 600]] + [1000000]
|
||||
|
||||
label, preliminary_submit_dataset_a = add_last_next_time4fault(label, preliminary_submit_dataset_a, TIME_INTERVAL,
|
||||
next_time_list)
|
||||
train, test = get_train_test(label, preliminary_submit_dataset_a, log_dataset)
|
||||
train = train.drop_duplicates(['sn', 'fault_time', 'time', 'msg', 'server_model']).reset_index(drop=True)
|
||||
|
||||
train['time_interval'] = (pd.to_datetime(train['fault_time']) - pd.to_datetime(train['time'])).apply(
|
||||
lambda x: x.total_seconds())
|
||||
test['time_interval'] = (pd.to_datetime(test['fault_time']) - pd.to_datetime(test['time'])).apply(
|
||||
lambda x: x.total_seconds())
|
||||
|
||||
all_data = pd.concat([train, test], axis=0, ignore_index=True)
|
||||
all_data = all_data.sort_values(['sn','server_model', 'fault_time', 'time'])
|
||||
w2v_feats = get_w2v_feats(all_data,
|
||||
f1_list = ['sn'],
|
||||
f2_list = ['msg_list', 'msg_0', 'msg_1', 'msg_2'])
|
||||
|
||||
# 获取 server_model_time_interval_stat_fea
|
||||
server_model_time_interval_stat_fea = get_server_model_time_interval_stat_fea(all_data)
|
||||
|
||||
msg_text_fea = get_msg_text_fea_all(all_data)
|
||||
# 获取时间差特征
|
||||
duration_minutes_fea = get_duration_minutes_fea(train, test)
|
||||
|
||||
# 获取时间server_model特征
|
||||
server_model_fea = get_server_model_fea(train, test)
|
||||
counter = get_word_counter(train)
|
||||
|
||||
# 获取时间 nearest_msg 特征
|
||||
nearest_msg_fea = get_nearest_msg_fea(train, test)
|
||||
# 获取时间 server_model beta_target 特征
|
||||
beta_target_fea = get_beta_target(train, test)
|
||||
|
||||
key = ['sn', 'fault_time', 'label', 'server_model']
|
||||
|
||||
fea_num = len(KEY_WORDS)
|
||||
time_list = [i * TIME_INTERVAL for i in next_time_list]
|
||||
train = get_feature(train, time_list, KEY_WORDS, fea_num, key=['sn', 'fault_time', 'label', 'server_model'])
|
||||
test = get_feature(test, time_list, KEY_WORDS, fea_num, key=['sn', 'fault_time', 'server_model'])
|
||||
|
||||
print('添加 时间差 特征')
|
||||
train = train.merge(duration_minutes_fea, on=['sn', 'fault_time', 'server_model'])
|
||||
test = test.merge(duration_minutes_fea, on=['sn', 'fault_time', 'server_model'])
|
||||
|
||||
print('添加 server_model特征')
|
||||
train = train.merge(server_model_fea, on=['sn', 'server_model'])
|
||||
test = test.merge(server_model_fea, on=['sn', 'server_model'])
|
||||
|
||||
print('添加 w2v_feats')
|
||||
train = train.merge(w2v_feats, on=['sn' ])
|
||||
test = test.merge(w2v_feats, on=['sn', ])
|
||||
|
||||
print('添加 nearest_msg 特征')
|
||||
train = train.merge(nearest_msg_fea, on=['sn', 'server_model', 'fault_time'])
|
||||
test = test.merge(nearest_msg_fea, on=['sn', 'server_model', 'fault_time'])
|
||||
|
||||
print('添加 beta_target 特征')
|
||||
train = train.merge(beta_target_fea, on=['sn', 'server_model', 'fault_time'])
|
||||
test = test.merge(beta_target_fea, on=['sn', 'server_model', 'fault_time'])
|
||||
|
||||
server_model_sn_fea_2 = get_server_model_sn_fea_2(train, test)
|
||||
print('添加 server_model_sn_fea_2 特征')
|
||||
train = train.merge(server_model_sn_fea_2, on=['sn', 'server_model'])
|
||||
test = test.merge(server_model_sn_fea_2, on=['sn', 'server_model'])
|
||||
|
||||
# crashdump_venus_fea = pd.read_csv(os.path.join(GENERATION_DIR,'crashdump_venus_fea.csv') )
|
||||
# print('添加 crashdump_venus_fea 特征')
|
||||
# print(train.shape,test.shape,crashdump_venus_fea.shape)
|
||||
# train = train.merge(crashdump_venus_fea, on=['sn' , 'fault_time'],how = 'left')
|
||||
# test = test.merge(crashdump_venus_fea, on=['sn', 'fault_time' ],how = 'left')
|
||||
# print(train.shape,test.shape )
|
||||
|
||||
crashdump_venus_fea = pd.read_csv(os.path.join(GENERATION_DIR,'crashdump_venus_fea_v1.csv') )
|
||||
print('添加 crashdump_venus_fea 特征')
|
||||
print(train.shape,test.shape,crashdump_venus_fea.shape)
|
||||
train = train.merge(crashdump_venus_fea, on=['sn' , 'fault_time'],how = 'left')
|
||||
test = test.merge(crashdump_venus_fea, on=['sn', 'fault_time' ],how = 'left')
|
||||
print(train.shape,test.shape )
|
||||
test.to_csv(os.path.join(GENERATION_DIR,'test.csv'),index =False)
|
||||
train.to_csv(os.path.join(GENERATION_DIR,'train.csv'),index =False)
|
||||
|
||||
# print('添加 msg_text_fea 特征')
|
||||
# train = train.merge(msg_text_fea, on=['sn', 'fault_time' ], how='left')
|
||||
# test = test.merge(msg_text_fea, on=['sn', 'fault_time'], how='left')
|
||||
|
||||
# print('添加 关键词交叉特征 ')
|
||||
# train,test = get_key_word_cross_fea(train,test)
|
||||
|
||||
# print('添加 server_model_time_interval_stat_fea 特征')
|
||||
# train = train.merge(server_model_time_interval_stat_fea, on=['server_model' ],how ='left')
|
||||
# test = test.merge(server_model_time_interval_stat_fea, on=['server_model' ],how ='left')
|
||||
|
||||
|
||||
use_less_cols_1 = ['last_last_msg_cnt', 'last_first_msg_cnt','time_diff_1_min',
|
||||
'last_msg_list_unique_LabelEnc', 'last_msg_0_unique_LabelEnc',
|
||||
'last_msg_1_unique_LabelEnc', 'last_msg_2_unique_LabelEnc',
|
||||
'last_msg_list_list_LabelEnc', 'last_msg_0_list_LabelEnc',
|
||||
'last_msg_1_list_LabelEnc', 'last_msg_2_list_LabelEnc',
|
||||
'last_msg_0_first_LabelEnc', 'last_msg_1_first_LabelEnc',
|
||||
'last_msg_2_first_LabelEnc', 'last_msg_0_last_LabelEnc',
|
||||
'last_msg_1_last_LabelEnc', 'last_msg_2_last_LabelEnc',
|
||||
'last_msg_last_LabelEnc', 'last_msg_first_LabelEnc']
|
||||
|
||||
use_less_col = [i for i in train.columns if train[i].nunique() < 2] + use_less_cols_1
|
||||
|
||||
|
||||
print(f'use_less_col:{len(use_less_col)}')
|
||||
use_cols = [i for i in train.columns if i not in ['sn', 'fault_time', 'label', 'server_model'] + use_less_col]
|
||||
cat_cols = ['server_model_LabelEnc', 'msg_LabelEnc', 'msg_0_LabelEnc', 'msg_1_LabelEnc', 'msg_2_LabelEnc',]
|
||||
use_cols = sorted(use_cols)
|
||||
print('使用的特征维度:',len(use_cols))
|
||||
|
||||
# cat_cols = []
|
||||
# for i in use_cols:
|
||||
# if '_LabelEnc' in i:
|
||||
# cat_cols.append(i)
|
||||
|
||||
oof_prob = np.zeros((train.shape[0], 4))
|
||||
test_prob = np.zeros((test.shape[0], 4))
|
||||
# seeds = [42,4242,40424,1024,2048]
|
||||
seeds = [42 ]
|
||||
for seed in seeds:
|
||||
oof_prob, test_prob, fea_imp_df, model_list = run_lgb(train[use_cols], train[['label']], test[use_cols], k=5,
|
||||
seed=seed, cat_cols=cat_cols)
|
||||
oof_prob +=oof_prob/len(seeds)
|
||||
test_prob +=test_prob/len(seeds)
|
||||
|
||||
weight = search_weight(train, train[['label']], oof_prob, init_weight=[1.0], class_num=4, step=0.001)
|
||||
oof_prob = oof_prob * np.array(weight)
|
||||
test_prob = test_prob * np.array(weight)
|
||||
|
||||
target_df = train[['sn', 'fault_time', 'label']]
|
||||
submit_df = train[['sn', 'fault_time']]
|
||||
submit_df['label'] = oof_prob.argmax(axis=1)
|
||||
|
||||
score = macro_f1(target_df=target_df, submit_df=submit_df)
|
||||
print(f'********************** BEST MACRO_F1 : {score} **********************')
|
||||
score = round(score, 5)
|
||||
|
||||
y_pred = test_prob.argmax(axis=1)
|
||||
result = test[['sn', 'fault_time']]
|
||||
result['label'] = y_pred
|
||||
result = preliminary_submit_dataset_a.merge(result, on=['sn', 'fault_time'], how='left')[['sn', 'fault_time', 'label']]
|
||||
result['label'] = result['label'].fillna(0).astype(int)
|
||||
|
||||
result.to_csv(os.path.join(RESULT_DIR, f'lgb_result.csv'), index=False)
|
||||
|
||||
fea_imp_df = fea_imp_df.reset_index(drop=True)
|
||||
fea_imp_df.to_csv(os.path.join(RESULT_DIR, f'./lgb_fea_imp_{int(score * 100000)}.csv'), index=False)
|
||||
|
||||
train_result_prob = pd.DataFrame(oof_prob).add_prefix('lgb_class_')
|
||||
test_result_prob = pd.DataFrame(test_prob).add_prefix('lgb_class_')
|
||||
train_result_prob['label'] = train['label']
|
||||
train_result_prob['sn'] = train['sn']
|
||||
train_result_prob['fault_time'] = train['fault_time']
|
||||
test_result_prob['sn'] = test['sn']
|
||||
test_result_prob['fault_time'] = test['fault_time']
|
||||
|
||||
result_prob = pd.concat([train_result_prob,test_result_prob],ignore_index = True)
|
||||
result_prob.to_csv(os.path.join(RESULT_DIR,f'lgb_prob_result.csv'),index = False)
|
||||
|
||||
|
||||
end_time = datetime.datetime.now()
|
||||
cost_time = end_time - start_time
|
||||
print('****************** LIGHTGBM COST TIME : ',str(cost_time),' ******************')
|
||||
|
||||
'''
|
||||
|
||||
v7 最优版本 线下 7356
|
||||
v8: v7 添加 关键词交叉特征 线下 0.7357 线上 7338
|
||||
v8.1 v7 添加 关键词交叉特征 并作为类别变量输入模型 0.73361
|
||||
v8.2 v7 添加 关键词交叉特征 并作为类别变量输入模型 删除 TOP_KEY_WORDS 7117
|
||||
v8.3 v7 添加 关键词交叉特征 并作为类别变量输入模型 使用 TOP_KEY_WORDS_2 7260
|
||||
v8.3 v7 添加 关键词交叉特征 并作为类别变量输入模型 添加 TOP_KEY_WORDS_2 7260
|
||||
|
||||
'''
|
||||
@ -0,0 +1,150 @@
|
||||
import warnings
|
||||
import datetime
|
||||
import lightgbm as lgb
|
||||
import numpy as np
|
||||
import pandas as pd
|
||||
from catboost import CatBoostClassifier
|
||||
from sklearn.model_selection import StratifiedKFold
|
||||
|
||||
from utils import N_ROUNDS
|
||||
import pickle
|
||||
import os
|
||||
warnings.filterwarnings('ignore')
|
||||
|
||||
|
||||
def get_model_feature_importances(model):
|
||||
feature_importances = pd.DataFrame()
|
||||
feature_importances['fea'] = model.feature_names_
|
||||
feature_importances['importances'] = model.feature_importances_
|
||||
feature_importances = feature_importances.sort_values('importances', ascending=False).reset_index(drop=True)
|
||||
|
||||
return feature_importances
|
||||
|
||||
|
||||
def run_cbt(train, target, test, k, seed, NUM_CLASS=4, cat_cols=[]):
|
||||
print('********************** RUN CATBOOST MODEL **********************')
|
||||
print(f'****************** 当前的 SEED {seed} ********************** ')
|
||||
folds = StratifiedKFold(n_splits=k, shuffle=True, random_state=seed)
|
||||
oof_prob = np.zeros((train.shape[0], NUM_CLASS))
|
||||
test_prob = np.zeros((test.shape[0], NUM_CLASS))
|
||||
feature_importance_df = []
|
||||
offline_score = []
|
||||
model_list = []
|
||||
|
||||
## K-Fold
|
||||
for fold, (trn_idx, val_idx) in enumerate(folds.split(train, target)):
|
||||
print("FOLD {} IS RUNNING...".format(fold + 1))
|
||||
trn_x, trn_y = train.loc[trn_idx], target.loc[trn_idx]
|
||||
val_x, val_y = train.loc[val_idx], target.loc[val_idx]
|
||||
catboost_model = CatBoostClassifier(
|
||||
iterations=N_ROUNDS,
|
||||
od_type='Iter',
|
||||
od_wait=120,
|
||||
max_depth=8,
|
||||
learning_rate=0.05,
|
||||
l2_leaf_reg=9,
|
||||
random_seed=seed,
|
||||
fold_len_multiplier=1.1,
|
||||
loss_function='MultiClass',
|
||||
logging_level='Verbose',
|
||||
# task_type="GPU"
|
||||
|
||||
)
|
||||
|
||||
start_time = datetime.datetime.now()
|
||||
|
||||
catboost_model.fit(trn_x,
|
||||
trn_y,
|
||||
eval_set=(val_x, val_y),
|
||||
use_best_model=True,
|
||||
verbose=800,
|
||||
early_stopping_rounds=100,
|
||||
cat_features=cat_cols,
|
||||
)
|
||||
end_time = datetime.datetime.now()
|
||||
model_train_cost_time = end_time - start_time
|
||||
print('****************** 模型训练 COST TIME : ',str(model_train_cost_time),' ******************')
|
||||
|
||||
start_time = datetime.datetime.now()
|
||||
oof_prob[val_idx] = catboost_model.predict_proba(train.loc[val_idx])
|
||||
end_time = datetime.datetime.now()
|
||||
model_pred_cost_time = end_time - start_time
|
||||
print('****************** 模型预测 COST TIME : ', str(model_pred_cost_time), ' ******************')
|
||||
# catboost_model = catboost_model.get_best_iteration()
|
||||
test_prob += catboost_model.predict_proba(test) / folds.n_splits
|
||||
print(catboost_model.get_best_score())
|
||||
offline_score.append(catboost_model.get_best_score()['validation']['MultiClass'])
|
||||
|
||||
feature_importance_df.append(get_model_feature_importances(catboost_model))
|
||||
model_list.append(catboost_model)
|
||||
with open(os.path.join('../model', f'cat_model_flod_{fold}.pkl'), 'wb') as f:
|
||||
pickle.dump(catboost_model, f)
|
||||
print('\nOOF-MEAN-ERROR score:%.6f, OOF-STD:%.6f' % (np.mean(offline_score), np.std(offline_score)))
|
||||
fea_imp_df = pd.concat(feature_importance_df, ignore_index=True).groupby('fea').agg(
|
||||
{'importances': 'mean'}).reset_index().sort_values('importances', ascending=False).reset_index(drop=True)
|
||||
|
||||
return oof_prob, test_prob, fea_imp_df, model_list
|
||||
|
||||
|
||||
def run_lgb(train, target, test, k, seed=42, NUM_CLASS=4, cat_cols=[]):
|
||||
# feats = [f for f in train.columns if f not in ['cust_no', 'label', 'I7', 'I9', 'B6']]
|
||||
# print('Current num of features:', len(feats))
|
||||
print(f'********************** RUN LGBM MODEL **********************')
|
||||
print(f'****************** 当前的 SEED {seed} ********************** ')
|
||||
cols_map = {j: i for i, j in enumerate(train.columns)}
|
||||
cat_cols = [cols_map[i] for i in cat_cols]
|
||||
train = train.rename(columns=cols_map)
|
||||
test = test.rename(columns=cols_map)
|
||||
folds = StratifiedKFold(n_splits=k, shuffle=True, random_state=seed)
|
||||
oof_prob = np.zeros((train.shape[0], NUM_CLASS))
|
||||
test_prob = np.zeros((test.shape[0], NUM_CLASS))
|
||||
fea_imp_df_list = []
|
||||
offline_score = []
|
||||
model_list = []
|
||||
## K-Fold
|
||||
for fold, (trn_idx, val_idx) in enumerate(folds.split(train, target)):
|
||||
params = {
|
||||
"objective": "multiclass",
|
||||
"num_class": NUM_CLASS,
|
||||
"learning_rate": 0.01,
|
||||
"max_depth": -1,
|
||||
"num_leaves": 32,
|
||||
"verbose": -1,
|
||||
"bagging_fraction": 0.8,
|
||||
"feature_fraction": 0.8,
|
||||
"seed": seed,
|
||||
'metric': 'multi_error'
|
||||
|
||||
}
|
||||
print("FOLD {} IS RUNNING...".format(fold + 1))
|
||||
trn_data = lgb.Dataset(train.loc[trn_idx], label=target.loc[trn_idx])
|
||||
val_data = lgb.Dataset(train.loc[val_idx], label=target.loc[val_idx])
|
||||
|
||||
# train
|
||||
params['seed'] = seed
|
||||
lgb_model = lgb.train(
|
||||
params,
|
||||
trn_data,
|
||||
num_boost_round=N_ROUNDS,
|
||||
valid_sets=[trn_data, val_data],
|
||||
early_stopping_rounds=100,
|
||||
verbose_eval=200,
|
||||
categorical_feature=cat_cols,
|
||||
|
||||
)
|
||||
# predict
|
||||
oof_prob[val_idx] = lgb_model.predict(train.loc[val_idx], num_iteration=lgb_model.best_iteration)
|
||||
test_prob += lgb_model.predict(test, num_iteration=lgb_model.best_iteration) / folds.n_splits
|
||||
offline_score.append(lgb_model.best_score['valid_1']['multi_error'])
|
||||
fea_imp = pd.DataFrame()
|
||||
fea_imp['feature_name'] = lgb_model.feature_name()
|
||||
fea_imp['importance'] = lgb_model.feature_importance()
|
||||
fea_imp['feature_name'] = fea_imp['feature_name'].map({str(cols_map[i]): i for i in cols_map})
|
||||
fea_imp = fea_imp.sort_values('importance', ascending=False)
|
||||
fea_imp_df_list.append(fea_imp)
|
||||
|
||||
model_list.append(lgb_model)
|
||||
print('\nOOF-MEAN-ERROR score:%.6f, OOF-STD:%.6f' % (np.mean(offline_score), np.std(offline_score)))
|
||||
fea_imp_df = pd.concat(fea_imp_df_list, ignore_index=True).groupby('feature_name').agg(
|
||||
{'importance': 'mean'}).reset_index().sort_values('importance', ascending=False)
|
||||
return oof_prob, test_prob, fea_imp_df, model_list
|
||||
@ -0,0 +1 @@
|
||||
scikit_learn==1.0.2
|
||||
@ -0,0 +1,43 @@
|
||||
import os
|
||||
import numpy as np
|
||||
import pandas as pd
|
||||
from utils import RESULT_DIR
|
||||
lgb_result = pd.read_csv(os.path.join(RESULT_DIR,'lgb_prob_result.csv'))
|
||||
lgb_result = lgb_result[lgb_result['label'].isnull()]
|
||||
print(lgb_result.columns)
|
||||
del lgb_result['label']
|
||||
|
||||
cat_result = pd.read_csv(os.path.join(RESULT_DIR,'cat_prob_result.csv'))
|
||||
cat_result = cat_result[cat_result['label'].isnull()]
|
||||
del cat_result['label']
|
||||
|
||||
# bert_result = pd.read_csv(os.path.join(RESULT_DIR,'bert_prob_result.csv'))
|
||||
|
||||
model_weight = {'lgb':0.2,'cat':0.8,'bert':0.2}
|
||||
print(f'MODEL WEIGHT: {model_weight}')
|
||||
# for i in ['bert_class_0', 'bert_class_1', 'bert_class_2','bert_class_3']:
|
||||
# bert_result[i] = bert_result[i]*model_weight['bert']
|
||||
|
||||
for i in ['cat_class_0', 'cat_class_1', 'cat_class_2', 'cat_class_3']:
|
||||
cat_result[i] = cat_result[i]*model_weight['cat']
|
||||
|
||||
for i in ['lgb_class_0', 'lgb_class_1', 'lgb_class_2', 'lgb_class_3']:
|
||||
lgb_result[i] = lgb_result[i]*model_weight['lgb']
|
||||
|
||||
result= lgb_result.merge(cat_result,on =['sn', 'fault_time'],how ='left' )
|
||||
|
||||
# result= bert_result.merge(cat_result,on =['sn', 'fault_time'],how ='left' )
|
||||
#
|
||||
# result['class_0'] =result.loc[:,['cat_class_0','bert_class_0']].sum(1)
|
||||
# result['class_1'] =result.loc[:,['cat_class_1','bert_class_0']].sum(1)
|
||||
# result['class_2'] =result.loc[:,['cat_class_2','bert_class_0']].sum(1)
|
||||
# result['class_3'] =result.loc[:,['cat_class_3','bert_class_0']].sum(1)
|
||||
|
||||
result['class_0'] =result.loc[:,['lgb_class_0','cat_class_0']].sum(1)
|
||||
result['class_1'] =result.loc[:,['lgb_class_1','cat_class_1']].sum(1)
|
||||
result['class_2'] =result.loc[:,['lgb_class_2','cat_class_2']].sum(1)
|
||||
result['class_3'] =result.loc[:,['lgb_class_3','cat_class_3']].sum(1)
|
||||
|
||||
result['label'] = np.argmax(result.loc[:,['class_0', 'class_1', 'class_2', 'class_3']].values,axis = 1)
|
||||
result = result[['sn', 'fault_time','label']]
|
||||
result.to_csv(os.path.join(RESULT_DIR,'stacking_result.csv'),index = False)
|
||||
@ -0,0 +1 @@
|
||||
https://tianchi.aliyun.com/competition/entrance/531947/information?lang=zh-cn
|
||||
@ -0,0 +1,6 @@
|
||||
# 创建镜像 并提交到你的镜像仓库
|
||||
rm -rf result.zip
|
||||
# built 镜像
|
||||
docker build -t [你的仓库地址]:[TAG] .
|
||||
# push 镜像
|
||||
docker push [你的仓库地址]:[TAG]
|
||||
@ -0,0 +1,106 @@
|
||||
use_less_col:335
|
||||
使用的特征维度: 1762
|
||||
********************** RUN CATBOOST MODEL **********************
|
||||
****************** 当前的 SEED 42 **********************
|
||||
FOLD 1 IS RUNNING...
|
||||
0: learn: 1.2939006 test: 1.2943096 best: 1.2943096 (0) total: 135ms remaining: 22m 27s
|
||||
800: learn: 0.2057676 test: 0.2778188 best: 0.2778188 (800) total: 1m 2s remaining: 11m 55s
|
||||
1600: learn: 0.1533318 test: 0.2698555 best: 0.2698522 (1599) total: 2m 4s remaining: 10m 54s
|
||||
Stopped by overfitting detector (100 iterations wait)
|
||||
|
||||
bestTest = 0.2677497192
|
||||
bestIteration = 2222
|
||||
|
||||
Shrink model to first 2223 iterations.
|
||||
{'learn': {'MultiClass': 0.12163532058790176}, 'validation': {'MultiClass': 0.26774971916097773}}
|
||||
FOLD 2 IS RUNNING...
|
||||
0: learn: 1.2947765 test: 1.2944610 best: 1.2944610 (0) total: 81.8ms remaining: 13m 38s
|
||||
800: learn: 0.2009925 test: 0.2969940 best: 0.2969940 (800) total: 1m 2s remaining: 11m 53s
|
||||
Stopped by overfitting detector (100 iterations wait)
|
||||
|
||||
bestTest = 0.2898436422
|
||||
bestIteration = 1413
|
||||
|
||||
Shrink model to first 1414 iterations.
|
||||
{'learn': {'MultiClass': 0.15671545706553627}, 'validation': {'MultiClass': 0.2898436422052235}}
|
||||
FOLD 3 IS RUNNING...
|
||||
0: learn: 1.2956904 test: 1.2979653 best: 1.2979653 (0) total: 83.6ms remaining: 13m 55s
|
||||
800: learn: 0.2010365 test: 0.3031897 best: 0.3031249 (796) total: 1m 2s remaining: 11m 56s
|
||||
1600: learn: 0.1521093 test: 0.2952955 best: 0.2952927 (1598) total: 2m 4s remaining: 10m 54s
|
||||
Stopped by overfitting detector (100 iterations wait)
|
||||
|
||||
bestTest = 0.2948664255
|
||||
bestIteration = 1799
|
||||
|
||||
Shrink model to first 1800 iterations.
|
||||
{'learn': {'MultiClass': 0.13764700334845772}, 'validation': {'MultiClass': 0.2948664254808659}}
|
||||
FOLD 4 IS RUNNING...
|
||||
0: learn: 1.2944941 test: 1.2931731 best: 1.2931731 (0) total: 83.8ms remaining: 13m 58s
|
||||
800: learn: 0.2055831 test: 0.2798750 best: 0.2798750 (800) total: 1m 2s remaining: 11m 54s
|
||||
1600: learn: 0.1555797 test: 0.2733073 best: 0.2732265 (1590) total: 2m 4s remaining: 10m 54s
|
||||
Stopped by overfitting detector (100 iterations wait)
|
||||
|
||||
bestTest = 0.2729804824
|
||||
bestIteration = 1672
|
||||
|
||||
Shrink model to first 1673 iterations.
|
||||
{'learn': {'MultiClass': 0.14819996336927216}, 'validation': {'MultiClass': 0.27298048242230794}}
|
||||
FOLD 5 IS RUNNING...
|
||||
0: learn: 1.2909100 test: 1.2914652 best: 1.2914652 (0) total: 86.9ms remaining: 14m 29s
|
||||
800: learn: 0.2014462 test: 0.2983963 best: 0.2983963 (800) total: 1m 2s remaining: 11m 55s
|
||||
1600: learn: 0.1523926 test: 0.2909189 best: 0.2907775 (1582) total: 2m 4s remaining: 10m 54s
|
||||
Stopped by overfitting detector (100 iterations wait)
|
||||
|
||||
bestTest = 0.2898741689
|
||||
bestIteration = 1887
|
||||
|
||||
Shrink model to first 1888 iterations.
|
||||
{'learn': {'MultiClass': 0.13391467495348316}, 'validation': {'MultiClass': 0.289874168865446}}
|
||||
|
||||
OOF-MEAN-ERROR score:0.283063, OOF-STD:0.010657
|
||||
Inint Score: 0.7240031522090993
|
||||
round: 1
|
||||
class:0, new_weight:1.01, f1 score: 0.7242893038330873
|
||||
class:0, new_weight:1.02, f1 score: 0.7244468658037289
|
||||
class:0, new_weight:1.03, f1 score: 0.7247189260435818
|
||||
class:0, new_weight:1.05, f1 score: 0.7247883133652404
|
||||
class:0, new_weight:1.06, f1 score: 0.7253074441711662
|
||||
class:0, new_weight:1.07, f1 score: 0.7255838308898628
|
||||
class:0, new_weight:1.09, f1 score: 0.7258591461588992
|
||||
class:0, new_weight:1.1, f1 score: 0.7263732069942956
|
||||
class:0, new_weight:1.11, f1 score: 0.7269810148203093
|
||||
class:0, new_weight:1.12, f1 score: 0.727085092104794
|
||||
class:0, new_weight:1.19, f1 score: 0.7275673332111965
|
||||
class:0, new_weight:1.2, f1 score: 0.7277300984468054
|
||||
class:0, new_weight:1.21, f1 score: 0.7300337938032027
|
||||
class:0, new_weight:1.22, f1 score: 0.7302916982856817
|
||||
class:0, new_weight:1.32, f1 score: 0.7302972834627351
|
||||
class:0, new_weight:1.33, f1 score: 0.7305212560605624
|
||||
class:0, new_weight:1.34, f1 score: 0.7307742905548762
|
||||
class:0, new_weight:1.3800000000000001, f1 score: 0.731115696618317
|
||||
class:0, new_weight:1.3900000000000001, f1 score: 0.7311341774607671
|
||||
class:0, new_weight:1.4000000000000001, f1 score: 0.7321211157346706
|
||||
class:0, new_weight:1.41, f1 score: 0.732530278451288
|
||||
class:0, new_weight:1.42, f1 score: 0.7326514907204666
|
||||
class:0, new_weight:1.43, f1 score: 0.7326655042252155
|
||||
class:0, new_weight:1.44, f1 score: 0.7340465325949609
|
||||
class:0, new_weight:1.45, f1 score: 0.7349701799847135
|
||||
class:2, new_weight:0.47000000000000003, f1 score: 0.7351355277520346
|
||||
class:2, new_weight:0.51, f1 score: 0.7352366908078052
|
||||
class:2, new_weight:0.52, f1 score: 0.7354704485017871
|
||||
class:2, new_weight:0.53, f1 score: 0.7356003615547112
|
||||
class:2, new_weight:0.54, f1 score: 0.7358162977063339
|
||||
class:2, new_weight:0.55, f1 score: 0.7360528528073605
|
||||
class:2, new_weight:0.6, f1 score: 0.7360930396635706
|
||||
class:2, new_weight:0.62, f1 score: 0.7361315695490319
|
||||
class:3, new_weight:0.77, f1 score: 0.736236509770795
|
||||
class:3, new_weight:0.79, f1 score: 0.7362861930960579
|
||||
class:3, new_weight:0.8, f1 score: 0.73637330491084
|
||||
class:3, new_weight:0.81, f1 score: 0.7364039172775363
|
||||
class:3, new_weight:0.8200000000000001, f1 score: 0.7365143106346561
|
||||
class:3, new_weight:0.8300000000000001, f1 score: 0.7366247783303799
|
||||
round: 2
|
||||
class:2, new_weight:0.55, f1 score: 0.7366811046538598
|
||||
round: 3
|
||||
********************** SEARCH BEST WEIGHT : [1.45, 1.0, 0.55, 0.8300000000000001] **********************
|
||||
********************** BEST MACRO_F1 : 0.7366811046538598 **********************
|
||||
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
@ -0,0 +1,12 @@
|
||||
Archive: model.zip
|
||||
creating: model/deberta-base/
|
||||
inflating: model/debert_model_v21_128_fs_flod_5.h5
|
||||
inflating: model/debert_model_v21_128_fs_flod_6.h5
|
||||
inflating: model/debert_model_v21_128_fs_flod_8.h5
|
||||
inflating: model/README.txt
|
||||
inflating: model/weight_cs6399_fold_8_v21_128_fs.npy
|
||||
inflating: model/weight_cs6558_fold_5_v21_128_fs.npy
|
||||
inflating: model/weight_cs6614_fold_6_v21_128_fs.npy
|
||||
inflating: model/weight_fs6138_fold_8_v21_128_fs.npy
|
||||
inflating: model/weight_fs6280_fold_5_v21_128_fs.npy
|
||||
inflating: model/weight_fs6359_fold_6_v21_128_fs.npy
|
||||
@ -0,0 +1,5 @@
|
||||
rm -rf model
|
||||
#unzip model.zip
|
||||
python3 code/get_crashdump_venus_fea.py
|
||||
python3 code/catboost_fs.py
|
||||
zip -j result.zip prediction_result/catboost_result.csv
|
||||
@ -0,0 +1 @@
|
||||
https://tianchi.aliyun.com/competition/entrance/531947/information?lang=zh-cn
|
||||
Binary file not shown.
Loading…
Reference in new issue