Разделы, фильтры и просмотр текстов в UI; опция «не хранить видео»

- таблица sections (дерево через parent_id), у видео section_id - API: CRUD /sections, фильтры /videos (поиск, раздел с потомками, метод), GET /videos/{id}/text, PATCH раздела, DELETE вместе с файлами - pipeline: keep_video=false удаляет mp4 после расшифровки, section_id - UI: панель разделов с вложенностью и счётчиками, поиск, фильтр по методу, диалоги полного текста/выжимки, привязка к разделу, удаление, переключатель «Сохранять видео на диске» - compose: bind-mount ./web для правок без пересборки Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-14 12:07:55 +05:00
parent 05ba753ade
commit bf854f2d6e
12 changed files with 612 additions and 43 deletions
--- a/api/route/pipeline.py
+++ b/api/route/pipeline.py
@@ -42,6 +42,8 @@ class ProcessRequest(BaseModel):
    url: str
    model_path: str | None = None  # path inside models/, used when subtitles are unavailable
    summary_max_sentences: int = 15
+    keep_video: bool = True  # False: файл видео удаляется после расшифровки
+    section_id: int | None = None
    # If supplied, this LLM is used for the summary; otherwise env-configured LLM;
    # otherwise the built-in extractive fallback.
    llm: LLMOverride | None = None
@@ -170,7 +172,14 @@ async def process_video(req: ProcessRequest):
        raise HTTPException(status_code=500, detail=f"Summarization failed: {e}")
    text_summary_path.write_text(summary, encoding="utf-8")

-    rel_video = str(video_path.relative_to(_PROJECT_ROOT))
+    if req.keep_video:
+        rel_video = str(video_path.relative_to(_PROJECT_ROOT))
+    else:
+        # пользователь просил не хранить видео: текст уже извлечён, файлы не нужны
+        video_path.unlink(missing_ok=True)
+        for p in _VIDEO_DIR.glob(f"{video_uuid}*.vtt"):
+            p.unlink(missing_ok=True)
+        rel_video = ""
    rel_full = str(text_full_path.relative_to(_PROJECT_ROOT))
    rel_summary = str(text_summary_path.relative_to(_PROJECT_ROOT))

@@ -183,6 +192,7 @@ async def process_video(req: ProcessRequest):
            text_full_path=rel_full,
            text_summary_path=rel_summary,
            transcription_method=method,
+            section_id=req.section_id,
        )
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"DB insert failed: {e}")
--- a/api/route/sections.py
+++ b/api/route/sections.py
@@ -0,0 +1,44 @@
+from fastapi import APIRouter, HTTPException
+from pydantic import BaseModel
+
+from api.db.section import create_section, delete_section, list_sections
+
+
+router = APIRouter(prefix="/sections", tags=["sections"])
+
+
+class SectionIn(BaseModel):
+    name: str
+    parent_id: int | None = None
+
+
+class SectionOut(BaseModel):
+    id: int
+    name: str
+    parent_id: int | None
+    video_count: int
+
+
+@router.get("/", response_model=list[SectionOut])
+async def list_all():
+    return await list_sections()
+
+
+@router.post("/", response_model=SectionOut)
+async def create(body: SectionIn):
+    name = body.name.strip()
+    if not name:
+        raise HTTPException(status_code=400, detail="Section name is empty")
+    try:
+        s = await create_section(name, body.parent_id)
+    except ValueError as e:
+        raise HTTPException(status_code=400, detail=str(e))
+    return SectionOut(id=s.id, name=s.name, parent_id=s.parent_id, video_count=0)
+
+
+@router.delete("/{section_id}")
+async def delete(section_id: int):
+    ok = await delete_section(section_id)
+    if not ok:
+        raise HTTPException(status_code=404, detail="Section not found")
+    return {"ok": True}
--- a/api/route/videos.py
+++ b/api/route/videos.py
@@ -1,18 +1,28 @@
 from datetime import datetime
-from typing import List
+from pathlib import Path
+from typing import List, Literal

-from fastapi import APIRouter, HTTPException
+from fastapi import APIRouter, HTTPException, Query
 from pydantic import BaseModel

-from api.db.video import delete_video, get_video, list_videos
+from api.db.section import list_sections
+from api.db.video import (
+    delete_video,
+    get_video,
+    list_videos,
+    set_video_section,
+)


 router = APIRouter(prefix="/videos", tags=["videos"])

+_PROJECT_ROOT = Path(__file__).resolve().parent.parent.parent
+

 class VideoOut(BaseModel):
    uuid: str
    source_url: str
+    section_id: int | None
    title: str
    video_path: str
    text_full_path: str
@@ -21,10 +31,15 @@ class VideoOut(BaseModel):
    created_at: datetime


+class SectionPatch(BaseModel):
+    section_id: int | None
+
+
 def _to_out(v) -> VideoOut:
    return VideoOut(
        uuid=v.uuid,
        source_url=v.source_url,
+        section_id=v.section_id,
        title=v.title,
        video_path=v.video_path,
        text_full_path=v.text_full_path,
@@ -34,9 +49,31 @@ def _to_out(v) -> VideoOut:
    )


+async def _descendant_ids(section_id: int) -> list[int]:
+    """Раздел + все его потомки (фильтр по разделу захватывает подразделы)."""
+    sections = await list_sections()
+    children: dict[int | None, list[int]] = {}
+    for s in sections:
+        children.setdefault(s["parent_id"], []).append(s["id"])
+    ids, stack = [], [section_id]
+    while stack:
+        cur = stack.pop()
+        ids.append(cur)
+        stack.extend(children.get(cur, []))
+    return ids
+
+
@router.get("/", response_model=List[VideoOut])
-async def list_all():
-    return [_to_out(v) for v in await list_videos()]
+async def list_all(
+    q: str | None = Query(default=None, description="Поиск по названию/URL"),
+    section_id: int | None = Query(default=None, description="Раздел (включая подразделы)"),
+    method: str | None = Query(default=None, description="Префикс метода: subtitles | vosk"),
+):
+    section_ids = await _descendant_ids(section_id) if section_id is not None else None
+    return [
+        _to_out(v)
+        for v in await list_videos(q=q, section_ids=section_ids, method=method)
+    ]


@router.get("/{video_uuid}", response_model=VideoOut)
@@ -47,9 +84,49 @@ async def get_one(video_uuid: str):
    return _to_out(v)


-@router.delete("/{video_uuid}")
-async def delete_one(video_uuid: str):
-    ok = await delete_video(video_uuid)
-    if not ok:
+@router.get("/{video_uuid}/text")
+async def get_text(video_uuid: str, kind: Literal["full", "summary"] = "full"):
+    v = await get_video(video_uuid)
+    if not v:
        raise HTTPException(status_code=404, detail="Video not found")
+    rel = v.text_full_path if kind == "full" else v.text_summary_path
+    if not rel:
+        raise HTTPException(status_code=404, detail="Text file is not recorded")
+    p = (_PROJECT_ROOT / rel).resolve()
+    try:
+        p.relative_to(_PROJECT_ROOT)
+    except ValueError:
+        raise HTTPException(status_code=400, detail="Bad text path")
+    if not p.exists():
+        raise HTTPException(status_code=404, detail=f"File missing on disk: {rel}")
+    return {"uuid": video_uuid, "kind": kind, "text": p.read_text(encoding="utf-8", errors="ignore")}
+
+
+@router.patch("/{video_uuid}", response_model=VideoOut)
+async def patch_section(video_uuid: str, body: SectionPatch):
+    v = await set_video_section(video_uuid, body.section_id)
+    if not v:
+        raise HTTPException(status_code=404, detail="Video not found")
+    return _to_out(v)
+
+
+@router.delete("/{video_uuid}")
+async def delete_one(video_uuid: str, delete_files: bool = True):
+    v = await get_video(video_uuid)
+    if not v:
+        raise HTTPException(status_code=404, detail="Video not found")
+    if delete_files:
+        for rel in (v.video_path, v.text_full_path, v.text_summary_path):
+            if not rel:
+                continue
+            p = (_PROJECT_ROOT / rel).resolve()
+            try:
+                p.relative_to(_PROJECT_ROOT)
+            except ValueError:
+                continue
+            p.unlink(missing_ok=True)
+        # сопутствующие субтитры .vtt рядом с видео
+        for p in (_PROJECT_ROOT / "data" / "videos").glob(f"{video_uuid}*.vtt"):
+            p.unlink(missing_ok=True)
+    await delete_video(video_uuid)
    return {"ok": True}