feat(tiktok): tiktok.scrape capability + billing wire-up

surfsense_backend/app/capabilities/core/billing.py

@@ -35,6 +35,7 @@ _PLATFORM_RATE_KEYS: dict[BillingUnit, str] = {
     BillingUnit.GOOGLE_MAPS_REVIEW: "GOOGLE_MAPS_MICROS_PER_REVIEW",
     BillingUnit.YOUTUBE_VIDEO: "YOUTUBE_MICROS_PER_VIDEO",
     BillingUnit.YOUTUBE_COMMENT: "YOUTUBE_MICROS_PER_COMMENT",
+    BillingUnit.TIKTOK_VIDEO: "TIKTOK_MICROS_PER_VIDEO",
 }
 
 
@@ -51,6 +52,7 @@ _UNIT_NOUNS: dict[BillingUnit, str] = {
     BillingUnit.GOOGLE_MAPS_REVIEW: "review",
     BillingUnit.YOUTUBE_VIDEO: "video",
     BillingUnit.YOUTUBE_COMMENT: "comment",
+    BillingUnit.TIKTOK_VIDEO: "video",
 }
 
 

surfsense_backend/app/capabilities/core/types.py

@@ -25,6 +25,7 @@ class BillingUnit(StrEnum):
     GOOGLE_MAPS_REVIEW = "google_maps_review"
     YOUTUBE_VIDEO = "youtube_video"
     YOUTUBE_COMMENT = "youtube_comment"
+    TIKTOK_VIDEO = "tiktok_video"
 
 
 class BillableInput(Protocol):

surfsense_backend/app/capabilities/tiktok/__init__.py

@@ -0,0 +1,5 @@
+"""``tiktok.*`` namespace: platform-native TikTok data verbs."""
+
+from __future__ import annotations
+
+from app.capabilities.tiktok.scrape import definition as _scrape  # noqa: F401

surfsense_backend/app/capabilities/tiktok/scrape/__init__.py

@@ -0,0 +1 @@
+"""``tiktok.scrape``: public TikTok videos over the browser-driven scraper."""

surfsense_backend/app/capabilities/tiktok/scrape/definition.py

@@ -0,0 +1,23 @@
+"""``tiktok.scrape`` capability registration (billed per video; see config
+``TIKTOK_MICROS_PER_VIDEO``)."""
+
+from __future__ import annotations
+
+from app.capabilities.core import BillingUnit, Capability, register_capability
+from app.capabilities.tiktok.scrape.executor import build_scrape_executor
+from app.capabilities.tiktok.scrape.schemas import ScrapeInput, ScrapeOutput
+
+TIKTOK_SCRAPE = Capability(
+    name="tiktok.scrape",
+    description=(
+        "Scrape public TikTok videos. Use urls, profiles, hashtags, or "
+        "search_queries."
+    ),
+    input_schema=ScrapeInput,
+    output_schema=ScrapeOutput,
+    executor=build_scrape_executor(),
+    billing_unit=BillingUnit.TIKTOK_VIDEO,
+    docs_url="/docs/connectors/native/tiktok",
+)
+
+register_capability(TIKTOK_SCRAPE)

surfsense_backend/app/capabilities/tiktok/scrape/executor.py

@@ -0,0 +1,48 @@
+"""``tiktok.scrape`` executor: verb input → scraper → TikTok video items."""
+
+from __future__ import annotations
+
+from collections.abc import Awaitable, Callable
+
+from app.capabilities.core import Executor
+from app.capabilities.core.progress import emit_progress
+from app.capabilities.tiktok.scrape.schemas import ScrapeInput, ScrapeOutput
+from app.exceptions import ForbiddenError
+from app.proprietary.platforms.tiktok import (
+    TikTokAccessBlockedError,
+    TikTokScrapeInput,
+    scrape_tiktok,
+)
+
+ScrapeFn = Callable[..., Awaitable[list[dict]]]
+
+
+def build_scrape_executor(scrape_fn: ScrapeFn | None = None) -> Executor:
+    """Bind the executor to a scraper fn (defaults to the proprietary actor)."""
+    scrape_fn = scrape_fn or scrape_tiktok
+
+    async def execute(payload: ScrapeInput) -> ScrapeOutput:
+        actor_input = TikTokScrapeInput(
+            startUrls=[{"url": url} for url in payload.urls],
+            profiles=payload.profiles,
+            hashtags=payload.hashtags,
+            searchQueries=payload.search_queries,
+            resultsPerPage=payload.results_per_page,
+        )
+        emit_progress(
+            "starting", "Resolving TikTok targets", total=payload.max_items, unit="item"
+        )
+        try:
+            items = await scrape_fn(actor_input, limit=payload.max_items)
+        except TikTokAccessBlockedError as exc:
+            # Anonymous-only scraper; a hard block can't be retried with creds.
+            raise ForbiddenError(
+                f"TikTok refused anonymous access: {exc}",
+                code="TIKTOK_ACCESS_BLOCKED",
+            ) from exc
+        emit_progress(
+            "done", f"Scraped {len(items)} item(s)", current=len(items), unit="item"
+        )
+        return ScrapeOutput(items=items)
+
+    return execute

surfsense_backend/app/capabilities/tiktok/scrape/schemas.py

@@ -0,0 +1,86 @@
+"""``tiktok.scrape`` I/O contracts.
+
+A lean, agent-friendly surface over ``TikTokScrapeInput``
+(``app/proprietary/platforms/tiktok``). The executor maps this to the full
+scraper input; the scraper's ``TikTokVideoItem`` is reused verbatim as the
+output element. Any TikTok URL kind (video, profile, hashtag, search) goes in
+``urls``; ``profiles``/``hashtags``/``search_queries`` are typed shortcuts.
+"""
+
+from __future__ import annotations
+
+from pydantic import BaseModel, Field, model_validator
+
+from app.proprietary.platforms.tiktok import TikTokVideoItem
+
+MAX_TIKTOK_SOURCES = 20
+"""Per-call cap on each source list: bounds a synchronous request's fan-out."""
+
+MAX_TIKTOK_ITEMS = 100
+"""Hard ceiling on items returned per call, regardless of the per-target count."""
+
+
+class ScrapeInput(BaseModel):
+    urls: list[str] = Field(
+        default_factory=list,
+        max_length=MAX_TIKTOK_SOURCES,
+        description=(
+            "TikTok URLs to scrape: a video, a profile (/@<user>), a hashtag "
+            "(/tag/<name>), or a search URL. Provide these OR profiles/hashtags/"
+            "search_queries (at least one source is required)."
+        ),
+    )
+    profiles: list[str] = Field(
+        default_factory=list,
+        max_length=MAX_TIKTOK_SOURCES,
+        description="Profile usernames (with or without a leading '@').",
+    )
+    hashtags: list[str] = Field(
+        default_factory=list,
+        max_length=MAX_TIKTOK_SOURCES,
+        description="Hashtag names to scrape, without the leading '#'.",
+    )
+    search_queries: list[str] = Field(
+        default_factory=list,
+        max_length=MAX_TIKTOK_SOURCES,
+        description="Search terms to run on TikTok.",
+    )
+    results_per_page: int = Field(
+        default=10,
+        ge=1,
+        le=MAX_TIKTOK_ITEMS,
+        description="Max videos to pull per profile/hashtag/search target.",
+    )
+    max_items: int = Field(
+        default=10,
+        ge=1,
+        le=MAX_TIKTOK_ITEMS,
+        description="Max total items to return across all sources.",
+    )
+
+    @model_validator(mode="after")
+    def _require_a_source(self) -> ScrapeInput:
+        if not any((self.urls, self.profiles, self.hashtags, self.search_queries)):
+            raise ValueError(
+                "Provide at least one of 'urls', 'profiles', 'hashtags', or "
+                "'search_queries'."
+            )
+        return self
+
+    @property
+    def estimated_units(self) -> int:
+        """Worst-case billable items for the pre-flight gate: ``max_items`` is a
+        hard cross-source ceiling (le=100), so no call can exceed it."""
+        return self.max_items
+
+
+class ScrapeOutput(BaseModel):
+    items: list[TikTokVideoItem] = Field(
+        default_factory=list,
+        description="One item per video returned, in emission order.",
+    )
+
+    @property
+    def billable_units(self) -> int:
+        """One returned item = one billable unit."""
+        return len(self.items)