feat(posts): extract + record external file-host links (Phase 3)

Capture off-platform links (mega/gdrive/mediafire/dropbox/pixeldrain) embedded
in post bodies so they're never silently dropped, and surface them in the post
view. The download worker (Phase 4) walks these rows.

- link_extract.py: pure extractor — <a href> + bare URLs, unwraps Patreon
  redirect shims, PRESERVES the full url incl. #fragment (mega's key), dedups.
  Reusable by every platform (runs off Post.description).
- external_link model + migration 0049: post_id/artist_id/host/url/label/status
  /attempts/last_error/attachment_id/timing; CHECK whitelists (full enum incl.
  worker statuses up front) + (post_id,url) unique.
- importer._sync_external_links: insert-missing on both import paths
  (_apply_sidecar + upsert_post_record) so a re-import never resets a link's
  status; runs for all platforms.
- post_feed_service.get_post: returns external_links (detail-only).
- PostCard: renders the links (host chip + label + status) once expanded.
- tests: extractor (5 hosts, fragment, shim unwrap, dedup), importer (record +
  no-dup on reimport), serializer.

Refs FC #830.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

tests/test_link_extract.py

@@ -0,0 +1,76 @@
+"""Unit tests for link_extract — pure, no DB/network."""
+
+from backend.app.services.link_extract import extract_external_links, host_for
+
+
+def test_host_classification_covers_all_and_rejects_others():
+    assert host_for("https://mega.nz/file/x#k") == "mega"
+    assert host_for("https://www.dropbox.com/s/x?dl=0") == "dropbox"
+    assert host_for("https://drive.google.com/file/d/ID/view") == "gdrive"
+    assert host_for("https://pixeldrain.com/u/abc") == "pixeldrain"
+    assert host_for("https://www.mediafire.com/file/x") == "mediafire"
+    assert host_for("https://example.com/x") is None
+
+
+def test_anchor_preserves_fragment_and_label():
+    html = ('<p>Watch: '
+            '<a href="https://mega.nz/file/Bux0FRrZ#b2Pn9841">Mega - Streamable</a></p>')
+    links = extract_external_links(html)
+    assert len(links) == 1
+    assert links[0].host == "mega"
+    # The #fragment (mega's decryption key) MUST survive.
+    assert links[0].url == "https://mega.nz/file/Bux0FRrZ#b2Pn9841"
+    assert links[0].label == "Mega - Streamable"
+
+
+def test_extracts_all_five_hosts():
+    html = (
+        '<a href="https://mega.nz/file/a#k">m</a>'
+        '<a href="https://drive.google.com/file/d/ID/view?usp=sharing">g</a>'
+        '<a href="https://www.mediafire.com/file/x/y">mf</a>'
+        '<a href="https://www.dropbox.com/s/x/y?dl=0">db</a>'
+        '<a href="https://pixeldrain.com/u/ems4UomN">px</a>'
+    )
+    hosts = {lnk.host for lnk in extract_external_links(html)}
+    assert hosts == {"mega", "gdrive", "mediafire", "dropbox", "pixeldrain"}
+
+
+def test_bare_url_in_text():
+    html = "Full film: https://pixeldrain.com/u/ems4UomN and more prose."
+    links = extract_external_links(html)
+    assert len(links) == 1
+    assert links[0].host == "pixeldrain"
+    assert links[0].url == "https://pixeldrain.com/u/ems4UomN"  # trailing '.' trimmed
+    assert links[0].label is None
+
+
+def test_unwraps_patreon_redirect():
+    html = ('<a href="https://www.patreon.com/redirect?'
+            'url=https%3A%2F%2Fmega.nz%2Ffile%2Fz%23key">x</a>')
+    links = extract_external_links(html)
+    assert len(links) == 1
+    assert links[0].host == "mega"
+    assert links[0].url == "https://mega.nz/file/z#key"
+
+
+def test_ignores_unsupported_and_dedups_anchor_label_wins():
+    html = (
+        '<a href="https://example.com/x">no</a>'
+        '<a href="https://mega.nz/file/dup#k">a</a>'
+        ' bare https://mega.nz/file/dup#k'
+    )
+    links = extract_external_links(html)
+    assert len(links) == 1            # example.com dropped; mega dup collapsed
+    assert links[0].label == "a"      # anchor (with label) wins over bare
+
+
+def test_amp_escaped_query_is_unescaped():
+    html = '<a href="https://drive.google.com/uc?id=ID&amp;export=download">g</a>'
+    links = extract_external_links(html)
+    assert links[0].url == "https://drive.google.com/uc?id=ID&export=download"
+
+
+def test_empty_inputs():
+    assert extract_external_links(None) == []
+    assert extract_external_links("") == []
+    assert extract_external_links("<p>no links here</p>") == []

tests/test_post_feed_service.py

@@ -11,6 +11,7 @@ from sqlalchemy import select
 
 from backend.app.models import (
     Artist,
+    ExternalLink,
     ImageProvenance,
     ImageRecord,
     Post,
@@ -437,6 +438,27 @@ async def test_get_post_returns_sanitized_html_body(db):
     assert "<script>" not in item["description_html"]
 
 
+@pytest.mark.asyncio
+async def test_get_post_returns_external_links(db):
+    artist = await _seed_artist(db, "alice-extlinks")
+    src = await _seed_source(db, artist.id, "patreon", "https://p/alice-extlinks")
+    p = await _seed_post(
+        db, src.id, external_id="EXTLINKS", post_date=datetime.now(UTC),
+    )
+    db.add(ExternalLink(
+        post_id=p.id, artist_id=artist.id, host="mega",
+        url="https://mega.nz/file/x#k", label="Mega",
+    ))
+    await db.commit()
+
+    item = await PostFeedService(db).get_post(p.id)
+    assert len(item["external_links"]) == 1
+    link = item["external_links"][0]
+    assert link["host"] == "mega"
+    assert link["url"] == "https://mega.nz/file/x#k"
+    assert link["status"] == "pending"
+
+
 @pytest.mark.asyncio
 async def test_get_post_unknown_returns_none(db):
     item = await PostFeedService(db).get_post(99999)

tests/test_sidecar_import.py

@@ -9,6 +9,7 @@ from sqlalchemy import func, select
 
 from backend.app.models import (
     Artist,
+    ExternalLink,
     ImageProvenance,
     ImageRecord,
     ImportSettings,
@@ -226,3 +227,47 @@ def test_upsert_post_record_idempotent_no_double(importer, import_layout):
     assert importer.session.execute(
         select(func.count()).select_from(Post)
     ).scalar_one() == 1
+
+
+def test_external_links_recorded_from_body(importer, import_layout):
+    """A mega/gdrive/etc. link in the body is recorded as an external_link row
+    (status pending) so it's never silently dropped."""
+    import_root, _ = import_layout
+    artist = Artist(name="Linker", slug="linker")
+    importer.session.add(artist)
+    importer.session.flush()
+    sc = import_root / "Linker" / "_post.json"
+    sc.parent.mkdir(parents=True, exist_ok=True)
+    sc.write_text(json.dumps({
+        "category": "patreon", "id": 999, "title": "Film",
+        "url": "https://patreon.com/posts/999",
+        "content": '<p>Get it: <a href="https://mega.nz/file/AbC#key123">Mega</a></p>',
+    }))
+    assert importer.upsert_post_record(sc, artist=artist) is True
+    importer.session.expire_all()  # re-read so the server_default status loads
+    links = importer.session.execute(select(ExternalLink)).scalars().all()
+    assert len(links) == 1
+    assert links[0].host == "mega"
+    assert links[0].url == "https://mega.nz/file/AbC#key123"
+    assert links[0].label == "Mega"
+    assert links[0].status == "pending"
+
+
+def test_external_links_not_duplicated_on_reimport(importer, import_layout):
+    """Re-importing the same body keeps ONE external_link row (insert-missing)."""
+    import_root, _ = import_layout
+    artist = Artist(name="Linker2", slug="linker2")
+    importer.session.add(artist)
+    importer.session.flush()
+    sc = import_root / "Linker2" / "_post.json"
+    sc.parent.mkdir(parents=True, exist_ok=True)
+    sc.write_text(json.dumps({
+        "category": "patreon", "id": 1000, "title": "Film",
+        "url": "https://patreon.com/posts/1000",
+        "content": '<a href="https://pixeldrain.com/u/xyz">px</a>',
+    }))
+    assert importer.upsert_post_record(sc, artist=artist) is True
+    assert importer.upsert_post_record(sc, artist=artist) is True
+    assert importer.session.execute(
+        select(func.count()).select_from(ExternalLink)
+    ).scalar_one() == 1