avmc 0.1.0__tar.gz

avmc-0.1.0/PKG-INFO

@@ -0,0 +1,105 @@
+Metadata-Version: 2.4
+Name: avmc
+Version: 0.1.0
+Summary: Standalone AV metadata capture tool
+Author: BossaMelon
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: requests>=2.31
+Requires-Dist: beautifulsoup4>=4.12
+Requires-Dist: lxml>=5.0
+Requires-Dist: Pillow>=10.0
+
+# AVMC
+
+A standalone AV metadata capture tool (current source: `javbus`) with a clean pipeline.
+
+## Features
+- Accepts a single video file or a directory path.
+- If a directory is provided, recursively scans video files and processes them one by one.
+- Detects subtitle-style filenames and appends `-C` to output names.
+- Organizes output as `success_output_folder/Actor1,Actor2,Actor3/Number` (or `success_output_folder/Number` when no actor).
+- Optional subtitle badge on poster image.
+- Uses local `config.json` by default; no dependency on parent project config.
+
+## Project Structure
+- `main.py`: CLI entry
+- `pipeline.py`: scan + process pipeline
+- `sources/javbus.py`: metadata scraper
+- `io_ops.py`: image download/crop/badge, NFO writing, file move
+- `config.json`: runtime config
+
+## Installation
+```bash
+cd avmc
+pip install -r requirements.txt
+```
+
+## Usage
+Run as module:
+```bash
+python -m avmc /path/to/video_or_dir
+```
+
+Run as script:
+```bash
+python avmc/main.py /path/to/video_or_dir
+```
+
+Options:
+- `-c, --config`: config file path (default: `avmc/config.json`)
+- `-p, --proxy`: temporary proxy override (higher priority than config/env)
+- `--debug`: dump raw HTML to `.adc_debug/`; keep source video in place and create a symlink in output
+
+## Config
+Default config file: `avmc/config.json`
+
+```json
+{
+  "success_output_folder": "output",
+  "failed": {
+    "move_enabled": false,
+    "output_folder": "failed"
+  },
+  "proxy": {
+    "proxy": "",
+    "timeout": 10,
+    "retry": 3
+  },
+  "javbus": {
+    "cookie": "existmag=all"
+  },
+  "scan": {
+    "escape_folders": ["output"]
+  },
+  "subtitle_badge": {
+    "enabled": true,
+    "backup_enabled": true
+  },
+  "image": {
+    "jpeg_quality": 85,
+    "optimize": true,
+    "progressive": true
+  }
+}
+```
+
+## Subtitle Detection
+Subtitle flag is inferred from filename patterns (case-insensitive), including compact forms like:
+- `ABC123C`
+- `ABC123CH`
+- `ABC123CHS`
+- `ABC123CHT`
+
+When detected:
+- output number becomes `NUMBER-C`
+- NFO adds `中文字幕` tag/genre
+- poster badge can be applied when enabled
+
+## Notes
+- If scraping fails, source video is kept in place by default.
+- Set `failed.move_enabled=true` to move failed files into `failed.output_folder`.
+- Image host may return 403 depending on network/proxy/cookie status.
+
+## Development
+- See `AGENTS.md` for coding-agent execution rules in this folder.

avmc-0.1.0/README.md

@@ -0,0 +1,93 @@
+# AVMC
+
+A standalone AV metadata capture tool (current source: `javbus`) with a clean pipeline.
+
+## Features
+- Accepts a single video file or a directory path.
+- If a directory is provided, recursively scans video files and processes them one by one.
+- Detects subtitle-style filenames and appends `-C` to output names.
+- Organizes output as `success_output_folder/Actor1,Actor2,Actor3/Number` (or `success_output_folder/Number` when no actor).
+- Optional subtitle badge on poster image.
+- Uses local `config.json` by default; no dependency on parent project config.
+
+## Project Structure
+- `main.py`: CLI entry
+- `pipeline.py`: scan + process pipeline
+- `sources/javbus.py`: metadata scraper
+- `io_ops.py`: image download/crop/badge, NFO writing, file move
+- `config.json`: runtime config
+
+## Installation
+```bash
+cd avmc
+pip install -r requirements.txt
+```
+
+## Usage
+Run as module:
+```bash
+python -m avmc /path/to/video_or_dir
+```
+
+Run as script:
+```bash
+python avmc/main.py /path/to/video_or_dir
+```
+
+Options:
+- `-c, --config`: config file path (default: `avmc/config.json`)
+- `-p, --proxy`: temporary proxy override (higher priority than config/env)
+- `--debug`: dump raw HTML to `.adc_debug/`; keep source video in place and create a symlink in output
+
+## Config
+Default config file: `avmc/config.json`
+
+```json
+{
+  "success_output_folder": "output",
+  "failed": {
+    "move_enabled": false,
+    "output_folder": "failed"
+  },
+  "proxy": {
+    "proxy": "",
+    "timeout": 10,
+    "retry": 3
+  },
+  "javbus": {
+    "cookie": "existmag=all"
+  },
+  "scan": {
+    "escape_folders": ["output"]
+  },
+  "subtitle_badge": {
+    "enabled": true,
+    "backup_enabled": true
+  },
+  "image": {
+    "jpeg_quality": 85,
+    "optimize": true,
+    "progressive": true
+  }
+}
+```
+
+## Subtitle Detection
+Subtitle flag is inferred from filename patterns (case-insensitive), including compact forms like:
+- `ABC123C`
+- `ABC123CH`
+- `ABC123CHS`
+- `ABC123CHT`
+
+When detected:
+- output number becomes `NUMBER-C`
+- NFO adds `中文字幕` tag/genre
+- poster badge can be applied when enabled
+
+## Notes
+- If scraping fails, source video is kept in place by default.
+- Set `failed.move_enabled=true` to move failed files into `failed.output_folder`.
+- Image host may return 403 depending on network/proxy/cookie status.
+
+## Development
+- See `AGENTS.md` for coding-agent execution rules in this folder.

avmc-0.1.0/__init__.py

@@ -0,0 +1,5 @@
+"""Clean implementation for AV Data Capture."""
+
+from .main import run
+
+__all__ = ["run"]

avmc-0.1.0/__main__.py

@@ -0,0 +1,5 @@
+from avmc.main import main
+
+
+if __name__ == "__main__":
+    main()

avmc-0.1.0/avmc.egg-info/PKG-INFO

@@ -0,0 +1,105 @@
+Metadata-Version: 2.4
+Name: avmc
+Version: 0.1.0
+Summary: Standalone AV metadata capture tool
+Author: BossaMelon
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: requests>=2.31
+Requires-Dist: beautifulsoup4>=4.12
+Requires-Dist: lxml>=5.0
+Requires-Dist: Pillow>=10.0
+
+# AVMC
+
+A standalone AV metadata capture tool (current source: `javbus`) with a clean pipeline.
+
+## Features
+- Accepts a single video file or a directory path.
+- If a directory is provided, recursively scans video files and processes them one by one.
+- Detects subtitle-style filenames and appends `-C` to output names.
+- Organizes output as `success_output_folder/Actor1,Actor2,Actor3/Number` (or `success_output_folder/Number` when no actor).
+- Optional subtitle badge on poster image.
+- Uses local `config.json` by default; no dependency on parent project config.
+
+## Project Structure
+- `main.py`: CLI entry
+- `pipeline.py`: scan + process pipeline
+- `sources/javbus.py`: metadata scraper
+- `io_ops.py`: image download/crop/badge, NFO writing, file move
+- `config.json`: runtime config
+
+## Installation
+```bash
+cd avmc
+pip install -r requirements.txt
+```
+
+## Usage
+Run as module:
+```bash
+python -m avmc /path/to/video_or_dir
+```
+
+Run as script:
+```bash
+python avmc/main.py /path/to/video_or_dir
+```
+
+Options:
+- `-c, --config`: config file path (default: `avmc/config.json`)
+- `-p, --proxy`: temporary proxy override (higher priority than config/env)
+- `--debug`: dump raw HTML to `.adc_debug/`; keep source video in place and create a symlink in output
+
+## Config
+Default config file: `avmc/config.json`
+
+```json
+{
+  "success_output_folder": "output",
+  "failed": {
+    "move_enabled": false,
+    "output_folder": "failed"
+  },
+  "proxy": {
+    "proxy": "",
+    "timeout": 10,
+    "retry": 3
+  },
+  "javbus": {
+    "cookie": "existmag=all"
+  },
+  "scan": {
+    "escape_folders": ["output"]
+  },
+  "subtitle_badge": {
+    "enabled": true,
+    "backup_enabled": true
+  },
+  "image": {
+    "jpeg_quality": 85,
+    "optimize": true,
+    "progressive": true
+  }
+}
+```
+
+## Subtitle Detection
+Subtitle flag is inferred from filename patterns (case-insensitive), including compact forms like:
+- `ABC123C`
+- `ABC123CH`
+- `ABC123CHS`
+- `ABC123CHT`
+
+When detected:
+- output number becomes `NUMBER-C`
+- NFO adds `中文字幕` tag/genre
+- poster badge can be applied when enabled
+
+## Notes
+- If scraping fails, source video is kept in place by default.
+- Set `failed.move_enabled=true` to move failed files into `failed.output_folder`.
+- Image host may return 403 depending on network/proxy/cookie status.
+
+## Development
+- See `AGENTS.md` for coding-agent execution rules in this folder.

avmc-0.1.0/avmc.egg-info/SOURCES.txt

@@ -0,0 +1,38 @@
+README.md
+__init__.py
+__main__.py
+config.json
+context.py
+http_client.py
+io_ops.py
+main.py
+models.py
+numbering.py
+pipeline.py
+pyproject.toml
+settings.py
+./__init__.py
+./__main__.py
+./config.json
+./context.py
+./http_client.py
+./io_ops.py
+./main.py
+./models.py
+./numbering.py
+./pipeline.py
+./settings.py
+avmc.egg-info/PKG-INFO
+avmc.egg-info/SOURCES.txt
+avmc.egg-info/dependency_links.txt
+avmc.egg-info/entry_points.txt
+avmc.egg-info/requires.txt
+avmc.egg-info/top_level.txt
+sources/__init__.py
+sources/base.py
+sources/javbus.py
+tests/test_io_ops.py
+tests/test_javbus.py
+tests/test_nfo.py
+tests/test_numbering.py
+tests/test_pipeline.py

avmc-0.1.0/avmc.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

avmc-0.1.0/avmc.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+avmc = avmc.main:main

avmc-0.1.0/avmc.egg-info/requires.txt

@@ -0,0 +1,4 @@
+requests>=2.31
+beautifulsoup4>=4.12
+lxml>=5.0
+Pillow>=10.0

avmc-0.1.0/avmc.egg-info/top_level.txt

@@ -0,0 +1 @@
+avmc

avmc-0.1.0/config.json

@@ -0,0 +1,29 @@
+{
+  "success_output_folder": "output",
+  "failed": {
+    "move_enabled": false,
+    "output_folder": "failed"
+  },
+  "proxy": {
+    "proxy": "",
+    "timeout": 10,
+    "retry": 3
+  },
+  "javbus": {
+    "cookie": "existmag=all"
+  },
+  "scan": {
+    "escape_folders": [
+      "output"
+    ]
+  },
+  "subtitle_badge": {
+    "enabled": true,
+    "backup_enabled": true
+  },
+  "image": {
+    "jpeg_quality": 85,
+    "optimize": true,
+    "progressive": true
+  }
+}

avmc-0.1.0/context.py

@@ -0,0 +1,38 @@
+from dataclasses import dataclass
+import os
+from pathlib import Path
+
+from .http_client import HttpClient
+from .settings import AppConfig
+
+
+@dataclass
+class AppContext:
+    config: AppConfig
+    http: HttpClient
+    debug: bool = False
+    debug_dir: Path = Path(".adc_debug")
+
+
+def build_context(config_path: str, proxy_override: str | None = None, debug: bool = False) -> AppContext:
+    conf = AppConfig.from_path(config_path)
+    proxy_from_config, timeout, retry = conf.proxy()
+
+    if proxy_override is not None:
+        proxy = proxy_override.strip()
+    else:
+        proxy = _env_https_proxy() or proxy_from_config
+
+    http = HttpClient(proxy=proxy, timeout=timeout, retry=retry)
+    ctx = AppContext(config=conf, http=http, debug=debug)
+    if ctx.debug:
+        ctx.debug_dir.mkdir(parents=True, exist_ok=True)
+    return ctx
+
+
+def _env_https_proxy() -> str | None:
+    value = os.getenv("https_proxy") or os.getenv("HTTPS_PROXY")
+    if value is None:
+        return None
+    value = value.strip()
+    return value if value else None

avmc-0.1.0/http_client.py

@@ -0,0 +1,60 @@
+import requests
+
+
+class HttpClient:
+    def __init__(self, proxy: str, timeout: int, retry: int):
+        self.proxy = proxy
+        self.timeout = timeout
+        self.retry = retry
+        self.proxies = _build_proxies(proxy)
+        self.session = requests.Session()
+
+    def get_text(self, url: str, headers: dict | None = None, cookies: dict | None = None) -> str:
+        resp = self.get_response(url, headers=headers, cookies=cookies)
+        resp.encoding = "utf-8"
+        return resp.text
+
+    def get_bytes(self, url: str, headers: dict | None = None, cookies: dict | None = None) -> bytes:
+        resp = self.get_response(url, headers=headers, cookies=cookies)
+        return resp.content
+
+    def get_response(self, url: str, headers: dict | None = None, cookies: dict | None = None) -> requests.Response:
+        return self._request("get", url, headers=headers, cookies=cookies)
+
+    def _request(
+        self,
+        method: str,
+        url: str,
+        headers: dict | None = None,
+        cookies: dict | None = None,
+        data: dict | None = None,
+    ) -> requests.Response:
+        last_error = None
+        for i in range(self.retry):
+            try:
+                resp = self.session.request(
+                    method=method,
+                    url=url,
+                    timeout=self.timeout,
+                    headers=headers,
+                    cookies=cookies,
+                    data=data,
+                    proxies=self.proxies,
+                )
+                resp.raise_for_status()
+                return resp
+            except requests.RequestException as err:
+                last_error = err
+                print(f"[-] Connect retry {i + 1}/{self.retry}: {err}")
+        raise RuntimeError(f"Request failed: {method.upper()} {url}") from last_error
+
+
+def _build_proxies(proxy: str | None) -> dict[str, str] | None:
+    if not proxy:
+        return None
+    proxy_value = proxy.strip()
+    if not proxy_value:
+        return None
+    if "://" not in proxy_value:
+        proxy_value = f"http://{proxy_value}"
+    return {"http": proxy_value, "https": proxy_value}