Bump version to v0.3.26 for release

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

@@ -31,10 +31,10 @@ jobs:
       fail-fast: false
       matrix:
         python: ['3.7', '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: "3.8"
+    
 python:
-  version: "3.8"
   install:
     - requirements: requirements.txt
     - requirements: dev-requirements.txt

Pipfile

@@ -4,8 +4,32 @@ verify_ssl = true
 name = "pypi"
 
 [packages]
-vk-api = "*"
+vk-api = ">=11.9.9"
-yt-dlp = "*"
+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"
 
 [dev-packages]
 sphinx-copybutton = "==0.5.0"

README.md

@@ -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

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

requirements.txt

@@ -5,14 +5,15 @@
 #    pipenv lock --requirements
 #
 
-brotli==1.0.9; platform_python_implementation == 'CPython'
+# -i https://pypi.org/simple
-certifi==2022.12.7; python_version >= '3.6'
+brotli>=1.0.9; platform_python_implementation >= 'CPython'
-charset-normalizer==3.0.1; python_version >= '3.6'
+certifi>=2022.12.7; python_version >= '3.6'
-idna==3.4; python_version >= '3.5'
+charset-normalizer>=3.0.1; python_version >= '3.6'
-mutagen==1.46.0; python_version >= '3.7'
+idna>=3.4; python_version >= '3.5'
-pycryptodomex==3.17; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'
+mutagen>=1.46.0; python_version >= '3.7'
-requests==2.28.2; python_version >= '3.7' and python_version < '4'
+pycryptodomex>=3.17; python_version >= '2.7' and python_version not in '3.0, 3.1, 3.2, 3.3, 3.4'
-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'
+requests>=2.28.2; python_version >= '3.7' and python_version < '4'
-vk-api==11.9.9
+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'
-websockets==10.4; python_version >= '3.7'
+vk-api>=11.9.9
-yt-dlp==2023.2.17
+websockets>=10.4; python_version >= '3.7'
+yt-dlp>=2023.2.17

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():
@@ -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", [""])
                         )
@@ -352,9 +353,10 @@ 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

@@ -15,7 +15,7 @@ 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)
 

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 = "12"
+_PATCH = "26"
 # 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 = ""