openmd-cli 0.1.0__py3-none-any.whl

openmd/backends/douyin_detail.py

@@ -0,0 +1,124 @@
+"""Bounded, transient Douyin single-work detail acquisition."""
+
+from __future__ import annotations
+
+import re
+from dataclasses import replace
+from datetime import datetime, timezone
+from urllib.parse import parse_qs, urlparse
+
+from openmd.backends.youtube_discovery import _run_opencli
+from openmd.core.detector import extract_first_url
+from openmd.core.models import ContentItem
+from openmd.core.opencli_output import load_opencli_json
+
+
+class DouyinDetailError(ValueError):
+    def __init__(self, code: str, summary: str):
+        super().__init__(summary)
+        self.code = code
+
+
+def douyin_author(payload: dict) -> str | None:
+    """Read nickname-shaped fields without stringifying API objects or IDs."""
+    fields = {str(key).lower().replace("_", ""): value for key, value in payload.items()}
+    for key in ("author", "nickname", "authorname", "user", "username", "userinfo"):
+        value = fields.get(key)
+        if isinstance(value, dict):
+            value = douyin_author(value) or value.get("name")
+        if isinstance(value, str) and value.strip():
+            return value.strip()
+    return None
+
+
+def douyin_published_at(payload: dict) -> str | None:
+    """Keep source calendar times; normalize Unix seconds/milliseconds in UTC."""
+    fields = {str(key).lower().replace("_", ""): value for key, value in payload.items()}
+    for key in ("publishedat", "createtime", "created", "createdat", "publishtime", "date"):
+        value = fields.get(key)
+        if not isinstance(value, (str, int, float)) or isinstance(value, bool):
+            continue
+        text = str(value).strip()
+        if not text:
+            continue
+        try:
+            if re.fullmatch(r"\d{8}", text):
+                return datetime.strptime(text, "%Y%m%d").date().isoformat()
+            if re.fullmatch(r"\d{10}(?:\.\d+)?|\d{13}(?:\.\d+)?", text):
+                seconds = float(text) / (1000 if len(text.split(".")[0]) == 13 else 1)
+                return datetime.fromtimestamp(seconds, tz=timezone.utc).isoformat()
+            datetime.fromisoformat(text.replace("Z", "+00:00"))
+            return text
+        except (ValueError, OverflowError, OSError):
+            continue
+    return None
+
+
+def douyin_video_id(value: str) -> str | None:
+    target = extract_first_url(value)
+    if re.fullmatch(r"\d{6,20}", target):
+        return target
+    parsed = urlparse(target if "://" in target else "https://" + target)
+    host = (parsed.hostname or "").lower()
+    if host != "douyin.com" and not host.endswith(".douyin.com") and host != "iesdouyin.com" and not host.endswith(".iesdouyin.com"):
+        return None
+    for key in ("modal_id", "aweme_id", "item_id"):
+        candidate = parse_qs(parsed.query).get(key, [""])[0]
+        if re.fullmatch(r"\d{6,20}", candidate):
+            return candidate
+    match = re.search(r"/(?:video|note)/(\d{6,20})(?:/|$)", parsed.path)
+    return match.group(1) if match else None
+
+
+class DouyinDetailProvider:
+    name = "opencli.douyin.detail"
+
+    def __init__(self, *, runner=None, timeout: int = 90):
+        self.runner = runner or _run_opencli
+        self.timeout = timeout
+
+    def resolve(self, item: ContentItem) -> ContentItem:
+        target = extract_first_url(item.url)
+        expected = douyin_video_id(target) or (item.id if item.id and item.id.isdigit() else None)
+        parsed = urlparse(target)
+        if not expected and (parsed.hostname or "").lower() not in {"v.douyin.com", "www.iesdouyin.com"}:
+            raise DouyinDetailError("douyin_invalid_video_target", "A video ID, video URL, modal URL, or video share link is required.")
+        command = ["opencli", "douyin", "detail", expected or target, "--window", "background", "--keep-tab", "false", "-f", "json"]
+        try:
+            code, stdout, _stderr = self.runner(command, timeout=self.timeout)
+        except Exception as exc:
+            raise DouyinDetailError("douyin_detail_read_failed", "The bounded video detail command could not complete.") from exc
+        if code:
+            raise DouyinDetailError("douyin_detail_read_failed", "The bounded video detail command failed; check the browser session or access permissions.")
+        try:
+            payload = load_opencli_json(stdout)
+            if isinstance(payload, list) and len(payload) == 1:
+                payload = payload[0]
+            if not isinstance(payload, dict) or payload.get("ok") is False:
+                raise ValueError("invalid detail")
+            video_id = str(payload.get("aweme_id") or payload.get("id") or "")
+            if not re.fullmatch(r"\d{6,20}", video_id):
+                raise ValueError("missing identity")
+        except (ValueError, TypeError) as exc:
+            raise DouyinDetailError("douyin_detail_invalid", "Video detail returned no valid single-work identity.") from exc
+        if expected and video_id != expected:
+            raise DouyinDetailError("douyin_detail_id_mismatch", "Video detail did not match the requested work ID.")
+        images = payload.get("images")
+        is_image = bool(images) or payload.get("media_kind") == "image_text" or payload.get("aweme_type") in {2, 68, 150}
+        raw_meta = {**item.raw_meta, "detail_resolved": True, "identity_unresolved": False, "content_id": video_id}
+        raw_meta.update({"duration": payload.get("duration_s", payload.get("duration", 0)), "media_kind": "image_text" if is_image else "video"})
+        if is_image:
+            # Image-post descriptions are author text, distinct from video
+            # captions. Keep an empty value too: a missing description must not
+            # fall back to an ID or a separately supplied display title.
+            raw_meta["author_text"] = payload.get("desc") if isinstance(payload.get("desc"), str) else ""
+            raw_meta["author_text_source"] = "detail_desc"
+        for key in ("play_url", "transcript_candidates", "subtitles", "caption_info", "captionInfo"):
+            if payload.get(key):
+                raw_meta[key] = payload[key]
+        # A video's description remains title metadata, never a transcript.
+        return replace(item, url=f"https://www.douyin.com/video/{video_id}", id=video_id,
+                       title=item.title if item.title and item.title != video_id else str(payload.get("desc") or payload.get("title") or video_id),
+                       author=item.author or douyin_author(payload),
+                       published_at=item.published_at or douyin_published_at(payload),
+                       source_type="image_text" if is_image else "video", raw_meta=raw_meta)