Skip to main content

Upgrade python project from paddle-1.x to paddle-2.0

Project description

Upgrade your python model from paddle-1.x to paddle-2.

Build Status Coverage Status Version License

Attention

paddle1to2 aims to convert python files from paddle-1.x to paddle-2 one by one, it won't handle indirect imports. e.g.

# filename "a.py"
import paddle.fluid as fluid
pass

# filename "b.py"
from a import *
class MyLayer(fluid.layers.Layer):
    pass

fluid.layers.Layer in "b.py" won't get converted. So you have to make sure you have import all used paddle modules, classes, objects directly for every python file before running paddle1to2.

Install

paddle1to2 support Linux, Mac OS, Windows(Git Bash is recommended), but it requires Python 3.5.4 or higher to run. Multi-Process is supported for Linux and Mac OS, Single-Process is support for Windows, this will lead to performance difference.

  1. install with pip
pip install -U paddle1to2
paddle1to2 --help # show help
paddle1to2 --inpath /path/to/model.py # upgrade your model from paddle-1.x to paddle-2.0

ATTENTION: If your device contains multiple versions of python, you may need to run the following commands instead:

python3 -m pip install -U paddle1to2
python3 -m paddle1to2 -h
python3 -m paddle1to2 --inpath /path/to/model.py
  1. install from source
git clone https://github.com/T8T9/paddle1to2.git
cd paddle1to2
python setup.py sdist bdist_wheel
pip install -U ./dist/paddle1to2-*.whl
paddle1to2 --help # show help
paddle1to2 --inpath /path/to/model.py # upgrade your model from paddle-1.x to paddle-2.0

Develop

If you are a develop, and you want to test your code quickly, you can run the following command in project directory:

python -m paddle1to2 --inpath /path/to/model.py

#or 

python paddle1to2/main.py --inpath /path/to/model.py

Moreover, if you want to run a specific refactor, you can use the following command:

python -m paddle1to2 --inpath /path/to/model.py --refactor <refactor_name>

use python -m paddle1to2 -h to see full list of all refactors.

if you want to run all unittest, use command:

python -m unittest discover paddle1to2/tests/
# or
python setup.py test

or use command:

python -m unittest paddle1to2/tests/test_refactor.py

to run specific test file.

Change Spec

change_spec is a python dict defined in spec.py, it defines the rules to refactor your code.

change_spec = {
    "path.to.old_api": {
        "alias": [
            "path.to.old_api_alias1",
            "path.to1.to2.old_api_alias2",
            ],
        "update_to": "path.to.new_api",
        "warning": "this api is deprecated.",
        "args_list": ["arg1", "arg2"],
        "args_change": [
                ["arg2", "arg2_rename"],
                ["arg3", ""],
                ["", "new_arg", "default_value"],
            ],
        "args_warning": {"arg1":"warning message"},
        "args_transformer": "_default_transformer",
    },
}
  • alias: a list of alias of main alias path.to.old_api, all alias will be replaced with main alias.
  • update_to: path.to.old_api will be replaced with this new api if specified.
  • warning: print specified warning message when path.to.old_api is found. This field will be ignored if update_to is specified.
  • args_list: is argument list of path.to.old_api.
  • args_change: a list of list. It contains following format:
    • ["arg", "new_arg"]: rename a argument, e.g. func(arg=value) -> func(new_arg=value)
    • ["arg", ""]: remove a argument, e.g. func(arg=value) -> func()
    • ["", "new_arg", "default_value"]: add a new argument, e.g. func(arg=value) -> func(arg=value, new_arg=default_value)
  • args_warning: print specified warning message for specified argument after apply args_change.
  • args_transformer: execute customized transformer on an AST node, it will be called after applying args_change to do further refactor.

Other Tools

  1. find pattern of specific code snippet, usage:
find_pattern 'import paddle'

find_pattern command will traverse all nodes in AST, if you see code snippet you want, type in 'y' to get pattern.

  1. find match node in specific code for specific pattern, usage:
find_match_node -ss 'import paddle' -ps 'any'

you can also specify "--print-results" option to got representation of matched node, specify "--print-lineno" to got line number of matched code.

Acknowledgements

  • Bowler: Safe code refactoring for modern Python projects.
  • lib2to3: A built-in python library to refactor python code.
  • fissix: A backport of latest lib2to3, with enhancements.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

paddle1to2-0.0.28.tar.gz (91.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

paddle1to2-0.0.28-py3-none-any.whl (108.1 kB view details)

Uploaded Python 3

File details

Details for the file paddle1to2-0.0.28.tar.gz.

File metadata

  • Download URL: paddle1to2-0.0.28.tar.gz
  • Upload date:
  • Size: 91.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.2.0 pkginfo/1.6.1 requests/2.24.0 setuptools/50.3.2 requests-toolbelt/0.9.1 tqdm/4.51.0 CPython/3.8.2

File hashes

Hashes for paddle1to2-0.0.28.tar.gz
Algorithm Hash digest
SHA256 ea0bea34f19115c153f07a83e380ce0c2346f630cab92e0de04642edb184fbb0
MD5 bfdf5720146b228d5ad5d16c9a4269d4
BLAKE2b-256 56f40b4cfbc35cfbf7b7888a4fffa4d6b4846e77072fcc153e5ae3d4df8b17bd

See more details on using hashes here.

File details

Details for the file paddle1to2-0.0.28-py3-none-any.whl.

File metadata

  • Download URL: paddle1to2-0.0.28-py3-none-any.whl
  • Upload date:
  • Size: 108.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.2.0 pkginfo/1.6.1 requests/2.24.0 setuptools/50.3.2 requests-toolbelt/0.9.1 tqdm/4.51.0 CPython/3.8.2

File hashes

Hashes for paddle1to2-0.0.28-py3-none-any.whl
Algorithm Hash digest
SHA256 8b3ae3920619b0ebbdfede6a5295c3cfe79a983e1e056033f11f72ce434e933c
MD5 02d9ed0b9e175b9a4c9b7948a72e0a8d
BLAKE2b-256 1f82ed5c41c1a77744024f88f470380de188d9b23677dffdcae57f49e4ac2757

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page