Stop For You from stamping OON before TES fills author_id - #18
Closed
Pitchfork-and-Torch wants to merge 19 commits into
Closed
Stop For You from stamping OON before TES fills author_id#18Pitchfork-and-Torch wants to merge 19 commits into
Pitchfork-and-Torch wants to merge 19 commits into
Conversation
in_network_ids is passed to the VF client without deduplication, while oon_ids is deduped four lines below. retweeted_tweet_id is pushed for every candidate that has one, so the same ID repeats once per retweet of a given post — most often when that post is going viral. Neither VfClient implementation dedupes its input: StratoVfClient builds one call per element, and XaiVfClient chunks by XAI_VF_MAX_BATCH_SIZE, so duplicates consume batch slots and can force an extra round trip. Not a correctness issue — results collapse into a HashMap keyed by tweet ID — but redundant work on the For You serving path. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Deduplicate in_network_ids before VF lookup
For You was classifying follow-graph membership before CoreData wrote author_id. TweetMixer often ships 0; that stamped OON and Recs-only VF drops then hit followed authors. Co-authored-by: Jon Bailey <Pitchfork-and-Torch@users.noreply.github.com>
cursor
Bot
force-pushed
the
cursor/in-network-after-core-data-f02e
branch
from
September 7, 2026 04:12
a5d787a to
cb9d627
Compare
Pitchfork-and-Torch
marked this pull request as ready for review
September 7, 2026 04:31
Owner
Author
|
Wrong base (fork PR). Ship vehicle is upstream xai-org#128. Per submit law: never PR Pitchfork-and-Torch/x-algorithm. Do not merge this. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Bug
For You (
phoenix_candidate_pipeline) runsInNetworkCandidateHydratorbeforeCoreDataCandidateHydrator. InNetwork readsauthor_idand writesin_networkonce. It never runs again.TweetMixer (and other thin sources) often ship
author_id = 0. InNetwork treats 0 as unfollowed and stampsin_network = Some(false). TES later writes the real author. The flag stays false.VFCandidateHydratorthen fetches that post atTimelineHomeRecommendations. Recs-only Drop rules fire: tweetDO_NOT_AMPLIFY,SPAM_HIGH_RECALL,NSFW_HIGH_RECALL,MALICIOUS_URL,FOSNR_ABUSE_INSULTS, and the rest of the Recs table. The same followed author on Home is Allow (or Interstitial for NSFW HP / gore / card). No user-visible label on the Recs Drop path.InNetworkCandidateHydratorbefore CoreData inphoenix_candidate_pipeline.rsVFCandidateHydrator(in_network.unwrap_or(false)→ Recs) thenVFFilterphoenix_scores_pipelinealready runs CoreData, then InNetworkThis is not rewriting Recs-only policy. User-label DNA still exempts followers. In-network retweet originals stay on Home. This is not xai-org#115 (Following quotes), xai-org#116 (OON RT originals), or xai-org#119 (RT wrapper primary id).
Fix
Run CoreData, then InNetwork, same as the scores pipeline. Bidirectional-follow stays where it is (not a VF input).
Tests
in_network = truefalseauthor_id = 0→false(the stale-OON stamp)false; TES-fill-then-stamp istrueStandalone decision-table harness (same predicate): 6 assertions passed.
cargo testcannot run here. Public dump has no Home Mixer manifest.Leftover
BidirectionalFollowHydratorstill runs before CoreData and also readsauthor_id. Separate from VF surface selection.Upstream port: this head is the Pitchfork dump branch. A clean cherry-pick of the four files can go to
xai-org/x-algorithm.