This commit is contained in:
2025-07-20 20:33:45 +04:00
parent 24e9a4896e
commit 90f9af3a38
4 changed files with 64396 additions and 1 deletions

63545
data/wb_reviews.csv Normal file

File diff suppressed because it is too large Load Diff

102
poetry.lock generated
View File

@@ -558,6 +558,30 @@ files = [
docs = ["ipython", "matplotlib", "numpydoc", "sphinx"] docs = ["ipython", "matplotlib", "numpydoc", "sphinx"]
tests = ["pytest", "pytest-cov", "pytest-xdist"] tests = ["pytest", "pytest-cov", "pytest-xdist"]
[[package]]
name = "dawg-python"
version = "0.7.2"
description = "Pure-python reader for DAWGs (DAFSAs) created by dawgdic C++ library or DAWG Python extension."
optional = false
python-versions = "*"
groups = ["main"]
files = [
{file = "DAWG-Python-0.7.2.tar.gz", hash = "sha256:4a5e3286e6261cca02f205cfd5516a7ab10190fa30c51c28d345808f595e3421"},
{file = "DAWG_Python-0.7.2-py2.py3-none-any.whl", hash = "sha256:4941d5df081b8d6fcb4597e073a9f60d5c1ccc9d17cd733e8744d7ecfec94ef3"},
]
[[package]]
name = "dawg2-python"
version = "0.9.0"
description = "Pure-python reader for DAWGs (DAFSAs) created by dawgdic C++ library or DAWG Python extension."
optional = false
python-versions = "<4.0,>=3.8"
groups = ["main"]
files = [
{file = "dawg2_python-0.9.0-py3-none-any.whl", hash = "sha256:4fab6fc097bd176cd783cd8421b757348ea5a460789e53b0f6bb64831380bab5"},
{file = "dawg2_python-0.9.0.tar.gz", hash = "sha256:adea0312acd1a958659e8448ce6899046c0858d0b6c8949a51eebdeb5a113e4a"},
]
[[package]] [[package]]
name = "debugpy" name = "debugpy"
version = "1.8.15" version = "1.8.15"
@@ -618,6 +642,17 @@ files = [
{file = "defusedxml-0.7.1.tar.gz", hash = "sha256:1bb3032db185915b62d7c6209c5a8792be6a32ab2fedacc84e01b52c51aa3e69"}, {file = "defusedxml-0.7.1.tar.gz", hash = "sha256:1bb3032db185915b62d7c6209c5a8792be6a32ab2fedacc84e01b52c51aa3e69"},
] ]
[[package]]
name = "docopt"
version = "0.6.2"
description = "Pythonic argument parser, that will make you smile"
optional = false
python-versions = "*"
groups = ["main"]
files = [
{file = "docopt-0.6.2.tar.gz", hash = "sha256:49b3a825280bd66b3aa83585ef59c4a8c82f2c8a522dbe754a8bc8d08c85c491"},
]
[[package]] [[package]]
name = "emoji" name = "emoji"
version = "2.14.1" version = "2.14.1"
@@ -2514,6 +2549,71 @@ files = [
[package.extras] [package.extras]
windows-terminal = ["colorama (>=0.4.6)"] windows-terminal = ["colorama (>=0.4.6)"]
[[package]]
name = "pymorphy2"
version = "0.9.1"
description = "Morphological analyzer (POS tagger + inflection engine) for Russian language."
optional = false
python-versions = "*"
groups = ["main"]
files = [
{file = "pymorphy2-0.9.1-py3-none-any.whl", hash = "sha256:a5224ec153cd6920a33961ae636793ad5185143bc8045b3f9a9f8bd62ecacf68"},
{file = "pymorphy2-0.9.1.tar.gz", hash = "sha256:86c447157dee2eb2341efbe4538e1281a754756ba1aa32da77a89614c58b560c"},
]
[package.dependencies]
dawg-python = ">=0.7.1"
docopt = ">=0.6"
pymorphy2-dicts-ru = ">=2.4,<3.0"
[package.extras]
fast = ["DAWG (>=0.8)"]
[[package]]
name = "pymorphy2-dicts-ru"
version = "2.4.417127.4579844"
description = "Russian dictionaries for pymorphy2"
optional = false
python-versions = "*"
groups = ["main"]
files = [
{file = "pymorphy2-dicts-ru-2.4.417127.4579844.tar.gz", hash = "sha256:78cad03adca605021abd387a3b2eb515c851b86e94682a1ef2354a2c74fcc196"},
{file = "pymorphy2_dicts_ru-2.4.417127.4579844-py2.py3-none-any.whl", hash = "sha256:9a322a6ee78fd4a5dceead0545c24b9a91687ad5df95cbac1b36f6c36cbb498a"},
]
[[package]]
name = "pymorphy3"
version = "2.0.4"
description = "Morphological analyzer (POS tagger + inflection engine) for Russian language."
optional = false
python-versions = "*"
groups = ["main"]
files = [
{file = "pymorphy3-2.0.4-py3-none-any.whl", hash = "sha256:bcda81d242b1217d625357724f144836f4da8fbcde8c760a170854463c1925ee"},
{file = "pymorphy3-2.0.4.tar.gz", hash = "sha256:31d7d6949d59424767caf82729cd8f231fbe9df2465c66465191b9c48deca790"},
]
[package.dependencies]
dawg2-python = ">=0.8.0"
pymorphy3-dicts-ru = "*"
setuptools = {version = ">=68.2.2", markers = "python_version >= \"3.12\""}
[package.extras]
cli = ["click"]
fast = ["DAWG2 (>=0.9.0,<1.0.0) ; platform_python_implementation == \"CPython\""]
[[package]]
name = "pymorphy3-dicts-ru"
version = "2.4.417150.4580142"
description = "Russian dictionaries for pymorphy2"
optional = false
python-versions = "*"
groups = ["main"]
files = [
{file = "pymorphy3-dicts-ru-2.4.417150.4580142.tar.gz", hash = "sha256:39ab379d4ca905bafed50f5afc3a3de6f9643605776fbcabc4d3088d4ed382b0"},
{file = "pymorphy3_dicts_ru-2.4.417150.4580142-py2.py3-none-any.whl", hash = "sha256:718bac64c73c10c16073a199402657283d9b64c04188b694f6d3e9b0d85440f4"},
]
[[package]] [[package]]
name = "pyparsing" name = "pyparsing"
version = "3.2.3" version = "3.2.3"
@@ -3584,4 +3684,4 @@ updater = ["alteryx-open-src-update-checker (>=3.1.0)"]
[metadata] [metadata]
lock-version = "2.1" lock-version = "2.1"
python-versions = "^3.12" python-versions = "^3.12"
content-hash = "70499691ada27acb13160ef23de5d4d31f31372efcf2c87b41ba039582110f33" content-hash = "9af04f5db7e8182f17ea04ae3c4ddf829bb73c2963572559eeeeecbe2a5ecc71"

View File

@@ -23,6 +23,8 @@ python-docx = "^1.1.2"
nltk = "^3.9.1" nltk = "^3.9.1"
emoji = "^2.14.1" emoji = "^2.14.1"
mahotas = "^1.4.18" mahotas = "^1.4.18"
pymorphy2 = "^0.9.1"
pymorphy3 = "^2.0.4"
[build-system] [build-system]
requires = ["poetry-core"] requires = ["poetry-core"]

748
vkr.ipynb Normal file

File diff suppressed because one or more lines are too long