reverting library dependencies

https://blog.readthedocs.com/use-build-os-config/

https://docs.readthedocs.io/en/stable/config-file/v2.html#build-tools-python

.github/dependabot.yml

@@ -1,11 +0,0 @@
-version: 2
-updates:
-- package-ecosystem: "pip"
-  directory: "/"
-  schedule:
-    interval: "daily"
-  open-pull-requests-limit: 10
-- package-ecosystem: "github-actions"
-  directory: "/"
-  schedule:
-    interval: "daily"

.github/workflows/main.yml

@@ -30,11 +30,11 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python: ['3.7', '3.10']
+        python: ['3.10']
-        task: # --show-capture=no on purpose
+        task: # --show-capture=no on purpose, -s for captchas
           - name: Test
             run: |
-              pytest --show-capture=no --color=yes tests/
+              pytest -s --show-capture=no --color=yes tests/
 
         include:
           - python: '3.10'

.readthedocs.yaml

@@ -4,8 +4,12 @@ sphinx:
   configuration: docs/source/conf.py
   fail_on_warning: false
 
+build:
+  os: "ubuntu-22.04"
+  tools:
+    python: "3.10"
+    
 python:
-  version: "3.8"
   install:
     - requirements: requirements.txt
     - requirements: dev-requirements.txt

Makefile

@@ -13,4 +13,4 @@ run-checks :
 	black .
 	flake8 .
 	mypy .
-	CUDA_VISIBLE_DEVICES='' pytest -v --color=yes --doctest-modules tests/ vk_url_scraper/
+	CUDA_VISIBLE_DEVICES='' pytest -v --color=yes .

Pipfile

@@ -4,8 +4,33 @@ verify_ssl = true
 name = "pypi"
 
 [packages]
+yt-dlp = ">=2023.2.17"
+flake8 = "*"
+mypy = ">=0.961"
+black = ">=22.3.0"
+isort = ">=5.10.1"
+pytest = "*"
+pytest-sphinx = "*"
+pytest-cov = "*"
+twine = ">=1.11.0"
+sphinx = "<5.1.0,>=4.3.0"
+furo = ">=2022.6.4.1"
+myst-parser = "<0.19.0,>=0.15.2"
+sphinx-copybutton = ">=0.5.0"
+sphinx-autobuild = ">=2021.3.14"
+sphinx-autodoc-typehints = "*"
+python-dotenv = ">=0.21.1"
+brotli = ">=1.0.9"
+certifi = ">=2023.7.22"
+charset-normalizer = ">=3.0.1"
+idna = ">=3.4"
+mutagen = ">=1.46.0"
+pycryptodomex = ">=3.17"
+requests = ">=2.28.2"
+urllib3 = ">=1.26.14"
+websockets = ">=10.4"
+# vk-api = {ref = "77b5a0d51a6bbf54d59554332f28a488615fbd6c", git = "git+https://github.com/python273/vk_api.git"}
 vk-api = "*"
-yt-dlp = "*"
 
 [dev-packages]
 sphinx-copybutton = "==0.5.0"

README.md

@@ -20,7 +20,7 @@ To use the library you will need a valid username/password combination for vk.co
 vk_url_scraper --help
 
 # scrape a URL and get the JSON result in the console
-vk_url_scraper -username "username here" --password "password here" --urls https://vk.com/wall12345_6789
+vk_url_scraper --username "username here" --password "password here" --urls https://vk.com/wall12345_6789
 # OR
 vk_url_scraper -u "username here" -p "password here" --urls https://vk.com/wall12345_6789
 # you can also have multiple urls
@@ -28,7 +28,7 @@ vk_url_scraper -u "username here" -p "password here" --urls https://vk.com/wall1
 
 # you can pass a token as well to avoid always authenticating 
 # and possibly getting captcha prompts
-# you can fetch the token from the bk_config.v2.json file generated under by searching for "access_token"
+# you can fetch the token from the vk_config.v2.json file generated under by searching for "access_token"
 vk_url_scraper -u "username" -p "password" -t "vktoken goes here" --urls https://vk.com/wall12345_6789
 
 # save the JSON output into a file
@@ -89,7 +89,7 @@ see [docs] for all available functions.
 2. To run all checks to `make run-checks` (fixes style) or individually
    1. To fix style: `black .` and `isort .` -> `flake8 .` to validate lint
    2. To do type checking: `mypy .`
-   3. To test: `pytest .` (`pytest -v --color=yes --doctest-modules tests/ vk_url_scraper/` to user verbose, colors, and test docstring examples)
+   3. To test: `pytest .` (`pytest -v --color=yes --doctest-modules tests/ vk_url_scraper/` to use verbose, colors, and test docstring examples)
 3. `make docs` to generate shpynx docs -> edit [config.py](docs/source/conf.py) if needed
 
 To test the command line interface available in [__main__.py](__vk_url_scraper/__main__.py) you need to pass the `-m` option to python like so: `python -m vk_url_scraper -u "" -p "" --urls ...`
@@ -97,10 +97,11 @@ To test the command line interface available in [__main__.py](__vk_url_scraper/_
 
 ## Releasing new version
 1. edit [version.py](vk_url_scraper/version.py) with proper versioning
-2. run `./scripts/release.sh` to create a tag and push, alternatively
+2. make sure to run `pipenv run pip freeze > requirements.txt` if you manage libs with pipenv
+3. run `./scripts/release.sh` to create a tag and push, alternatively
    1. `git tag vx.y.z` to tag version
    2. `git push origin vx.y.z` -> this will trigger workflow and put project on [pypi](https://pypi.org/project/vk-url-scraper/)
-3. go to https://readthedocs.org/ to deploy new docs version (if webhook is not setup)
+4. go to https://readthedocs.org/ to deploy new docs version (if webhook is not setup)
 
 ### Fixing a failed release
 

dev-requirements.txt

@@ -2,11 +2,11 @@
 flake8
 
 # Static type checking
-mypy==0.961
+mypy>=0.961
 
 # Automatic code formatting
-black==22.3.0
+black>=22.3.0
-isort==5.10.1
+isort>=5.10.1
 
 # Running tests
 pytest
@@ -24,19 +24,20 @@ wheel
 Sphinx>=4.3.0,<5.1.0
 
 # Sphinx theme: https://sphinx-themes.org/sample-sites/furo/
-furo==2022.6.4.1
+furo>=2022.6.4.1
 
 # Lets Sphinx parse markdown files in addition to rst.
 myst-parser>=0.15.2,<0.19.0
 
 # Adds a copy button to code examples in the docs.
-sphinx-copybutton==0.5.0
+sphinx-copybutton>=0.5.0
 
 # Live rebuilding and reloading of docs for developing locally.
-sphinx-autobuild==2021.3.14
+sphinx-autobuild>=2021.3.14
 
 # Automatically adds types to docs
 sphinx-autodoc-typehints
 
 # For parsing and comparing version numbers.
 packaging
+python-dotenv>=0.21.1

docs/source/installation.md

@@ -1,7 +1,7 @@
 Installation
 ============
 
-**vk-url-scraper** supports Python >= 3.7.
+**vk-url-scraper** supports Python >= 3.10.
 
 ## Installing with `pip`
 

requirements.txt

@@ -1,95 +1,94 @@
-#
+aiohttp==3.9.1
-# These requirements were autogenerated by pipenv
+aiosignal==1.3.1
-# To regenerate from the project's Pipfile, run:
+alabaster==0.7.16
-#
+async-timeout==4.0.3
-#    pipenv lock --requirements --dev
+attrs==23.2.0
-#
+Babel==2.14.0
-
+beautifulsoup4==4.12.3
-# Note: in pipenv 2020.x, "--dev" changed to emit both default and development
+black==24.1a1
-# requirements. To emit only development requirements, pass "--dev-only".
+bleach==6.0.0
-
+Brotli==1.1.0
-# -i https://pypi.org/simple
+certifi==2023.11.17
-alabaster==0.7.13; python_version >= '3.6'
+cffi==1.16.0
-attrs==22.2.0; python_version >= '3.6'
+charset-normalizer==3.3.2
-babel==2.11.0; python_version >= '3.6'
+click==8.1.7
-beautifulsoup4==4.11.2; python_version >= '3.6'
+colorama==0.4.6
-black==22.3.0
+commonmark==0.9.1
-bleach==6.0.0; python_version >= '3.7'
+coverage==7.4.0
-brotli==1.0.9; platform_python_implementation == 'CPython'
+cryptography==42.0.0
-certifi==2022.12.7; python_version >= '3.6'
+docutils==0.18.1
-cffi==1.15.1
+exceptiongroup==1.2.0
-charset-normalizer==3.0.1; python_version >= '3.6'
+flake8==7.0.0
-click==8.1.3; python_version >= '3.7'
+frozenlist==1.4.1
-colorama==0.4.6; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6'
+furo==2023.3.27
-coverage[toml]==7.2.0; python_version >= '3.7'
+idna==3.6
-cryptography==39.0.1; python_version >= '3.6'
+imagesize==1.4.1
-docutils==0.18.1; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'
+importlib-metadata==7.0.1
-exceptiongroup==1.1.0; python_version < '3.11'
+iniconfig==2.0.0
-flake8==6.0.0
+isort==6.0.0b2
-furo==2022.6.21
+jaraco.classes==3.3.0
-idna==3.4; python_version >= '3.5'
+jeepney==0.8.0
-imagesize==1.4.1; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
+Jinja2==3.1.3
-importlib-metadata==6.0.0; python_version >= '3.7'
+keyring==24.3.0
-iniconfig==2.0.0; python_version >= '3.7'
-isort==5.10.1
-jaraco.classes==3.2.3; python_version >= '3.7'
-jeepney==0.8.0; sys_platform == 'linux'
-jinja2==3.1.2; python_version >= '3.7'
-keyring==23.13.1; python_version >= '3.7'
 livereload==2.6.3
-markdown-it-py==2.2.0; python_version >= '3.7'
+markdown-it-py==2.2.0
-markupsafe==2.1.2; python_version >= '3.7'
+MarkupSafe==2.1.4
-mccabe==0.7.0; python_version >= '3.6'
+mccabe==0.7.0
-mdit-py-plugins==0.3.4; python_version >= '3.7'
+mdit-py-plugins==0.3.5
-mdurl==0.1.2; python_version >= '3.7'
+mdurl==0.1.2
-more-itertools==9.0.0; python_version >= '3.7'
+more-itertools==10.2.0
-mutagen==1.46.0; python_version >= '3.7'
+multidict==6.0.4
-mypy-extensions==1.0.0; python_version >= '3.5'
+mutagen==1.47.0
-mypy==0.961
+mypy==1.8.0
+mypy-extensions==1.0.0
 myst-parser==0.18.1
-packaging==23.0; python_version >= '3.7'
+nh3==0.2.15
-pathspec==0.11.0; python_version >= '3.7'
+packaging==23.2
-pkginfo==1.9.6; python_version >= '3.6'
+pathspec==0.12.1
-platformdirs==3.0.0; python_version >= '3.7'
+pkginfo==1.9.6
-pluggy==1.0.0; python_version >= '3.6'
+platformdirs==4.1.0
-pycodestyle==2.10.0; python_version >= '3.6'
+pluggy==1.3.0
+py==1.11.0
+pycodestyle==2.11.1
 pycparser==2.21
-pycryptodomex==3.17; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'
+pycryptodomex==3.20.0
-pyflakes==3.0.1; python_version >= '3.6'
+pyflakes==3.2.0
-pygments==2.14.0; python_version >= '3.6'
+Pygments==2.17.2
-pytest-cov==4.0.0
+pyparsing==3.0.9
+pytest==8.0.0rc2
+pytest-cov==4.1.0
 pytest-sphinx==0.5.0
-pytest==7.2.1
+python-dotenv==1.0.1
-python-dotenv==0.21.1
+pytz==2022.1
-pytz==2022.7.1
+PyYAML==6.0.1
-pyyaml==6.0; python_version >= '3.6'
+readme-renderer==42.0
-readme-renderer==37.3; python_version >= '3.7'
+requests==2.31.0
-requests-toolbelt==0.10.1; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
+requests-toolbelt==1.0.0
-requests==2.28.2; python_version >= '3.7' and python_version < '4'
+rfc3986==2.0.0
-rfc3986==2.0.0; python_version >= '3.7'
+rich==13.7.0
-rich==13.3.1; python_version >= '3.7'
+SecretStorage==3.3.3
-secretstorage==3.3.3; sys_platform == 'linux'
+six==1.16.0
-six==1.16.0; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3'
 snowballstemmer==2.2.0
-soupsieve==2.4; python_version >= '3.7'
+soupsieve==2.5
+Sphinx==5.0.2
 sphinx-autobuild==2021.3.14
 sphinx-autodoc-typehints==1.19.1
-sphinx-basic-ng==1.0.0b1; python_version >= '3.7'
+sphinx-basic-ng==1.0.0b2
-sphinx-copybutton==0.5.0
+sphinx-copybutton==0.5.2
-sphinx==5.0.2
+sphinxcontrib-applehelp==1.0.8
-sphinxcontrib-applehelp==1.0.4; python_version >= '3.8'
+sphinxcontrib-devhelp==1.0.6
-sphinxcontrib-devhelp==1.0.2; python_version >= '3.5'
+sphinxcontrib-htmlhelp==2.0.5
-sphinxcontrib-htmlhelp==2.0.1; python_version >= '3.8'
+sphinxcontrib-jsmath==1.0.1
-sphinxcontrib-jsmath==1.0.1; python_version >= '3.5'
+sphinxcontrib-qthelp==1.0.7
-sphinxcontrib-qthelp==1.0.3; python_version >= '3.5'
+sphinxcontrib-serializinghtml==1.1.10
-sphinxcontrib-serializinghtml==1.1.5; python_version >= '3.5'
+tomli==2.0.1
-tomli==2.0.1; python_version < '3.11'
+tornado==6.4
-tornado==6.2; python_version > '2.7'
 twine==4.0.2
-typing-extensions==4.5.0; python_version >= '3.7'
+typing_extensions==4.9.0
-urllib3==1.26.14; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4, 3.5'
+urllib3==2.1.0
-vk-api==11.9.9
+vk-api @ git+https://github.com/python273/vk_api.git@77b5a0d51a6bbf54d59554332f28a488615fbd6c
 webencodings==0.5.1
-websockets==10.4; python_version >= '3.7'
+websockets==12.0
-yt-dlp==2023.2.17
+yarl==1.9.4
-zipp==3.14.0; python_version >= '3.7'
+yt-dlp==2024.1.22.232713.dev0
+zipp==3.17.0

setup.py

@@ -57,7 +57,7 @@ setup(
     package_data={"vk_url_scraper": ["py.typed"]},
     install_requires=read_requirements("requirements.txt"),
     extras_require={"dev": read_requirements("dev-requirements.txt")},
-    python_requires=">=3.7",
+    python_requires=">=3.10",
     entry_points={
         "console_scripts": [
             "vk_url_scraper=vk_url_scraper.__main__:main",

tests/scraper_test.py

@@ -14,15 +14,16 @@ def test_login_fail():
         VkScraper("invalid", "combination")
 
 
-def test_login_custom_file():
+# disabled due to CI
-    session_filename = "test-session.json"
+# def test_login_custom_file():
-    VkScraper(
+#     session_filename = "test-session.json"
-        os.environ["VK_USERNAME"],
+#     VkScraper(
-        os.environ["VK_PASSWORD"],
+#         os.environ["VK_USERNAME"],
-        session_file=session_filename,
+#         os.environ["VK_PASSWORD"],
-    )
+#         session_file=session_filename,
-    assert os.path.isfile(session_filename)
+#     )
-    os.unlink(session_filename)
+#     assert os.path.isfile(session_filename)
+#     os.unlink(session_filename)
 
 
 def test_login_success():
@@ -80,7 +81,7 @@ def test_scrape_wall_url_with_photos():
         == "Хабаровск\nАллея героев\nПомолимся об укокоении воинов:\nАлександра, Игоря, Эдуарда, \nДионисия, Евгения, Александра, Артемия, Иннокентия, Андрея."
     )
     assert str(res[0]["datetime"]) == str(datetime.datetime(2022, 6, 15, 10, 37, 24))
-    assert len(res[0]["payload"]) == 17
+    assert len(res[0]["payload"]) == 18
     assert len(res[0]["attachments"].keys()) == 1
     assert list(res[0]["attachments"].keys()) == ["photo"]
     assert len(res[0]["attachments"]["photo"]) == 9
@@ -92,7 +93,7 @@ def test_scrape_wall_url_with_photos_inner_videos_and_links_with_inner_photos():
     assert res[0]["id"] == "wall-17315087_74182"
     assert res[0]["text"] == ""
     assert str(res[0]["datetime"]) == str(datetime.datetime(2022, 3, 24, 11, 1, 9))
-    assert len(res[0]["payload"]) == 17
+    assert len(res[0]["payload"]) == 18
     assert len(res[0]["attachments"].keys()) == 3
     for k in ["photo", "link", "video"]:
         assert k in list(res[0]["attachments"].keys())
@@ -127,7 +128,7 @@ def test_scrape_photo_only():
         == "Делимся расписанием конкурса [https://vk.com/wall-1_399468|«Код Петербурга»]. Все важные этапы — на одной схеме \n\nЕсли участвуете, обязательно сохраните себе. Так будет удобнее планировать работу над проектом, и вы точно не упустите лучший момент для отправки сервиса на модерацию."
     )
     assert str(res[0]["datetime"]) == str(datetime.datetime(2022, 6, 7, 9, 43))
-    assert len(res[0]["payload"]) == 15
+    assert len(res[0]["payload"]) == 16
     assert len(res[0]["attachments"].keys()) == 1
     assert list(res[0]["attachments"].keys()) == ["photo"]
     assert len(res[0]["attachments"]["photo"]) == 1
@@ -138,7 +139,7 @@ def test_scrape_video_only():
     assert len(res) == 1
     assert res[0]["id"] == "video38556806_456251917"
     assert str(res[0]["datetime"]) == str(datetime.datetime(2022, 3, 24, 5, 42, 38))
-    assert len(res[0]["payload"]) == 31
+    assert len(res[0]["payload"]) == 34
     assert len(res[0]["attachments"].keys()) == 1
     assert list(res[0]["attachments"].keys()) == ["video"]
 
@@ -149,3 +150,21 @@ def test_scrape_video_only2():
         vks.download_media(res, tempdir)
         found_files = set(os.listdir(tempdir))
         assert "video-17546758_456239898_0.mp4" in found_files
+
+
+def test_scrape_private_video():
+    """
+    > Some videos are kept private and cannot be accessed without a passkey . In this case, send the ID in {owner_id}_{video_id}_{access_key}.
+    From https://dev.vk.com/ru/method/video.get
+    """
+    res = vks.scrape("https://vk.com/wall-127774884_178565")
+
+    with tempfile.TemporaryDirectory(dir="./") as tempdir:
+        vks.download_media(res, tempdir)
+        expect_files = {
+            "wall-127774884_178565_0.mp4",
+            "wall-127774884_178565_1.mp4",
+            "wall-127774884_178565_2.mp4",
+        }
+        found_files = set(os.listdir(tempdir))
+        assert len(expect_files) == len(expect_files & found_files)

vk_url_scraper/__main__.py

@@ -19,7 +19,7 @@ def get_argument_parser():
         action="store",
         dest="username",
         required=True,
-        help="username for a valid vk.com account",
+        help="username for a valid vk.com account (pass empty if using --token)",
     )
     parser.add_argument(
         "-p",
@@ -27,7 +27,7 @@ def get_argument_parser():
         action="store",
         dest="password",
         required=True,
-        help="password for the valid vk.com account",
+        help="password for the valid vk.com account (pass empty if using --token)",
     )
     parser.add_argument(
         "-t",

vk_url_scraper/scraper.py

@@ -3,7 +3,7 @@ import re
 import shutil
 from collections import defaultdict
 from datetime import datetime
-from typing import List
+from typing import List, Optional
 from urllib.parse import urlparse
 
 import requests
@@ -37,13 +37,13 @@ class VkScraper:
 
     WALL_PATTERN = re.compile(r"(wall.{0,1}\d+_\d+)")
     PHOTO_PATTERN = re.compile(r"(photo.{0,1}\d+_\d+)")
-    VIDEO_PATTERN = re.compile(r"(video.{0,1}\d+_\d+)")
+    VIDEO_PATTERN = re.compile(r"(video.{0,1}\d+_\d+(?:_\w+)?)")
 
     def __init__(
         self,
         username: str,
         password: str,
-        token: str = None,
+        token: Optional[str] = None,
         session_file="vk_config.v2.json",
         captcha_handler=captcha_handler,
     ) -> None:
@@ -144,10 +144,11 @@ class VkScraper:
                     first_type = a["type"]
                     attachment = a[first_type]
                     if first_type == "video":
+                        video_path = f'video{attachment["owner_id"]}_{attachment["id"]}'
+                        if "access_key" in attachment:
+                            video_path += f"_{attachment['access_key']}"
                         attachments["video"].extend(
-                            self.scrape_videos(f'video{attachment["owner_id"]}_{attachment["id"]}')[
+                            self.scrape_videos(video_path)[0]
-                                0
-                            ]
                             .get("attachments", {})
                             .get("video", [""])
                         )
@@ -338,7 +339,9 @@ class VkScraper:
                             filename = os.path.join(destination, f"{r['id']}_{i}.%(ext)s")
                             ydl = yt_dlp.YoutubeDL(
                                 {
-                                    "format": "bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best",
+                                    "format": (
+                                        "bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best"
+                                    ),
                                     "merge_output_format": "mp4",
                                     "retries": 5,
                                     "noplaylist": True,
@@ -352,9 +355,10 @@ class VkScraper:
                             info = ydl.extract_info(url, download=True)
                             filename = ydl.prepare_filename(info)
                             if "unknown_video" in filename:
+                                old_filename = filename
                                 filename = shutil.copy(
-                                    filename, filename.replace("unknown_video", "mkv")
+                                    filename, filename.replace("unknown_video", "mp4")
                                 )
-                                os.remove(filename)
+                                os.remove(old_filename)
                             downloaded.append(filename)
         return downloaded

vk_url_scraper/utils.py

@@ -15,9 +15,9 @@ class DateTimeEncoder(json.JSONEncoder):
 
 def captcha_handler(captcha):
     key = input(
-        f"CAPTCHA DETECTED, please solve it and input the solution. url={captcha.get_url()}:"
+        f"CAPTCHA DETECTED, please solve it and input the solution. url= {captcha.get_url()} :"
     ).strip()
-    return captcha.try_again(key)
+    return captcha.try_again(key.strip())
 
 
 @contextmanager

vk_url_scraper/version.py

@@ -2,7 +2,7 @@ _MAJOR = "0"
 _MINOR = "3"
 # On main and in a nightly release the patch should be one ahead of the last
 # released build.
-_PATCH = "13"
+_PATCH = "28"
 # This is mainly for nightly builds which have the suffix ".dev$DATE". See
 # https://semver.org/#is-v123-a-semantic-version for the semantics.
 _SUFFIX = ""