sawz/last30days-skill
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: surface YouTube + TikTok top comments alongside Reddit (#260)

Browse Source 
* feat(normalize): pass YouTube top_comments through with Reddit-compatible shape

_normalize_youtube silently dropped top_comments after enrich_with_comments
populated them, so the downstream signals/render/entity layers never saw
YouTube comments. Map likes->score and text->excerpt so the existing
Reddit-compatible readers Just Work.

Shared _remap_comments helper will be reused for TikTok in a later commit.

* feat(tiktok): fetch top comments via ScrapeCreators when opted in

Mirrors the youtube_comments pattern: new env.is_tiktok_comments_available
gate (requires SCRAPECREATORS_API_KEY + tiktok_comments in INCLUDE_SOURCES),
tiktok.enrich_with_comments ranks posts and fetches via
GET /v1/tiktok/video/comments. Vote field is digg_count; text and user.nickname
come across verbatim. Pipeline calls the enricher right after TikTok search
when the gate is open.

Comment-fetch errors never crash the pipeline — the enricher returns an
empty list on 4xx/5xx.

* feat(normalize): pass TikTok top_comments through with digg_count->score mapping

Instagram uses the same shortform normalizer and has no comment fetcher
today, so the key is harmlessly absent there — no Instagram regression.

* feat(signals): add YouTube + TikTok top-comment score to engagement formula

Mirrors Reddit's 10% top-comment slot. Without top_comments present, the
formula reduces to views-dominant weighting; with a high-signal comment,
the item gets a meaningful bump (log1p(10k) ~ 9.2, weighted 0.10 = ~0.92
on the engagement score).

Updated the existing dominant-weight and missing-fields tests to the new
weights (0.45/0.32/0.13 for YT, 0.45/0.27/0.18 for TT). Views still dominate.

* feat(render): source-aware thresholds and vote labels for top comments

10 upvotes on Reddit signals community interest; 10 likes on a viral
TikTok is noise. Introduce per-source minimums (reddit 10, youtube 50,
tiktok 500) and native vote labels ('upvotes' for Reddit, 'likes' for
YT/TT). First-pass numbers — tune after live observation.

* docs: generalize top-comment quoting to YouTube + TikTok, add tiktok_comments opt-in

Synthesis instructions previously called out Reddit top comments only.
Now cover Reddit/YouTube/TikTok uniformly with source-appropriate vote
labels (upvotes vs likes), and explicitly frame YT transcript highlights
and comments as complementary signals. README and setup-wizard copy
document the new tiktok_comments INCLUDE_SOURCES token.

---------

Co-authored-by: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>
This commit is contained in:
Matt Van Horn
2026-04-15 08:26:06 -04:00
committed by GitHub
parent 242e38ef56
commit 082efe03e3
12 changed files with 678 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_signals_v3.py
+102 -9
View File
@@ -28,6 +28,98 @@ class SignalsV3Tests(unittest.TestCase):
)
self.assertAlmostEqual(expected, signals.engagement_raw(item))
def test_youtube_engagement_adds_top_comment_slot(self):
with_comment = schema.SourceItem(
item_id="yt1",
source="youtube",
title="Title",
body="Body",
url="https://youtube.com/watch?v=a",
engagement={"views": 10000, "likes": 500, "comments": 30},
metadata={"top_comments": [{"score": 500}]},
)
without = schema.SourceItem(
item_id="yt2",
source="youtube",
title="Title",
body="Body",
url="https://youtube.com/watch?v=b",
engagement={"views": 10000, "likes": 500, "comments": 30},
metadata={"top_comments": []},
)
with_score = signals.engagement_raw(with_comment)
without_score = signals.engagement_raw(without)
self.assertIsNotNone(with_score)
self.assertIsNotNone(without_score)
self.assertGreater(with_score, without_score)
expected = (
0.45 * math.log1p(10000)
+ 0.32 * math.log1p(500)
+ 0.13 * math.log1p(30)
+ 0.10 * math.log1p(500)
)
self.assertAlmostEqual(expected, with_score, places=6)
def test_youtube_engagement_empty_returns_none(self):
item = schema.SourceItem(
item_id="yt-empty",
source="youtube",
title="Title",
body="Body",
url="https://youtube.com/watch?v=e",
engagement={},
metadata={"top_comments": []},
)
self.assertIsNone(signals.engagement_raw(item))
def test_tiktok_engagement_adds_top_comment_slot(self):
item = schema.SourceItem(
item_id="tt1",
source="tiktok",
title="Title",
body="Body",
url="https://tiktok.com/@u/video/1",
engagement={"views": 100000, "likes": 5000, "comments": 500},
metadata={"top_comments": [{"score": 1200}]},
)
expected = (
0.45 * math.log1p(100000)
+ 0.27 * math.log1p(5000)
+ 0.18 * math.log1p(500)
+ 0.10 * math.log1p(1200)
)
self.assertAlmostEqual(expected, signals.engagement_raw(item), places=6)
def test_youtube_ranking_promotes_viral_comment_thread(self):
"""A moderately-viewed YouTube video with a 10k-like comment should
outrank a slightly-higher-viewed video with no high-signal comments."""
viral_comment = schema.SourceItem(
item_id="yt-with-viral-comment",
source="youtube",
title="Deploy to Fly.io",
body="Deploy to Fly.io walkthrough",
url="https://youtube.com/watch?v=x",
published_at="2026-03-15",
engagement={"views": 5000, "likes": 200, "comments": 50},
metadata={"top_comments": [{"score": 10000}]},
)
higher_views = schema.SourceItem(
item_id="yt-higher-views-no-comment",
source="youtube",
title="Deploy to Fly.io",
body="Deploy to Fly.io walkthrough",
url="https://youtube.com/watch?v=y",
published_at="2026-03-15",
engagement={"views": 8000, "likes": 300, "comments": 60},
metadata={"top_comments": []},
)
ranked = signals.annotate_stream(
[higher_views, viral_comment],
ranking_query="How do I deploy on Fly.io?",
freshness_mode="balanced_recent",
)
self.assertEqual("yt-with-viral-comment", ranked[0].item_id)
def test_polymarket_engagement_uses_market_fields(self):
item = schema.SourceItem(
item_id="pm1",
@@ -221,7 +313,8 @@ class SignalsV3Tests(unittest.TestCase):
self.assertAlmostEqual(expected, result)
def test_youtube_engagement_dominant_weight(self):
"""YouTube: views at 0.50 should dominate over comments at 0.15."""
"""YouTube: views at 0.45 should dominate. With no top-comment data,
the remaining 0.90 of weight is split views/likes/comments 0.45/0.32/0.13."""
item = schema.SourceItem(
item_id="yt1", source="youtube", title="T", body="B",
url="https://example.com",
@@ -230,9 +323,9 @@ class SignalsV3Tests(unittest.TestCase):
result = signals.engagement_raw(item)
self.assertIsNotNone(result)
expected = (
0.50 * math.log1p(10000)
+ 0.35 * math.log1p(500)
+ 0.15 * math.log1p(80)
0.45 * math.log1p(10000)
+ 0.32 * math.log1p(500)
+ 0.13 * math.log1p(80)
)
self.assertAlmostEqual(expected, result)
@@ -252,7 +345,7 @@ class SignalsV3Tests(unittest.TestCase):
)
result = signals.engagement_raw(item)
self.assertIsNotNone(result)
expected = 0.50 * math.log1p(5000)
expected = 0.45 * math.log1p(5000)
self.assertAlmostEqual(expected, result)
def test_tiktok_engagement_dominant_weight(self):
@@ -264,9 +357,9 @@ class SignalsV3Tests(unittest.TestCase):
result = signals.engagement_raw(item)
self.assertIsNotNone(result)
expected = (
0.50 * math.log1p(50000)
+ 0.30 * math.log1p(3000)
+ 0.20 * math.log1p(200)
0.45 * math.log1p(50000)
+ 0.27 * math.log1p(3000)
+ 0.18 * math.log1p(200)
)
self.assertAlmostEqual(expected, result)
@@ -286,7 +379,7 @@ class SignalsV3Tests(unittest.TestCase):
)
result = signals.engagement_raw(item)
self.assertIsNotNone(result)
expected = 0.30 * math.log1p(1000)
expected = 0.27 * math.log1p(1000)
self.assertAlmostEqual(expected, result)
def test_instagram_engagement_dominant_weight(self):