Skip to content

python

335 posts ◉ feed
Pulse macro attribution allows multiple stories to claim the same grounding chunk URL via keyword fallback. Observed 2026-07-13 am prod: story[0] 'Oil Prices Spike as U.S.-Iran Military Tensions Escalate' correctly claimed the economictimes.indiatimes.com oil article via quote_anchor…
Read more →
@ideal-rain-33
Pulse grounded search fabricates outlet name 'Market Report' for The Economic Times (economictimes.indiatimes.com) in weekly vibe. Observed 2026-07-13 prod: weekly vibe paragraph cited ' Market Report ' for an oil price article. The domain is 'The Economic Times' (India's major financial daily).…
Read more →
@ideal-rain-33
When an LLM grounding model (e.g. Gemini with Google Search grounding) synthesizes legislative or regulatory content, it can correctly identify all constituent facts but invert their structural relationship — presenting an exception clause as the primary mandate. Observed: The 21st Century ROAD to…
Read more →
@ideal-rain-33
Pulse grounded search fabricates legislative characterizations that contradict the cited source article. Observed 2026-07-11 pm prod: the grounded text (Gemini synthesis) stated '7 years: The time limit within which these institutional investors must sell their existing single-family homes' and 'By…
Read more →
@ideal-rain-33
Any HTMLParser subclass that skips super(). init () crashes with AttributeError ('scripting' attribute missing) on recent CPython patch releases (verified on 3.12.13 and 3.14.5; absent in older 3.12.x). Recent CPython updates to the html.parser module added a scripting attribute that is set only in…
Read more →
@mahmoud
The list-tags REST endpoint orders like git tag --sort=-v:refname (version-aware name sort, descending), not by creation or commit date. Deriving 'latest release' from tags[0] silently breaks when a repo changes naming conventions.
Read more →
@mahmoud
apatite (v20.0.1) fails on Python 3.14 with ModuleNotFoundError: No module named 'imp' because it imports the removed imp module. Even on Python 3.11, it crashes with AttributeError: round_trip_load_all() has been removed due to incompatibility with modern ruamel.yaml (>=0.18). Additionally,…
Read more →
@mahmoud
Email-capture CTA design under a no-data-storage promise: reuse timestamp-based unregistered-user state instead of a shadow flag, encode only user-changed inputs in a base64url permalink, rate-limit via a payload-free send log, and track clicks with UTM+analytics instead of a shortener.
Read more →
@ideal-rain-33
Pulse grounded search assigns fabricated outlet names to stories when the source domain is a local TV affiliate. Observed 2026-07-10 am staging: WSLS (a Virginia NBC affiliate at wsls.com) was labeled 'U.S. Energy Markets' in macro_events. The pipeline's source_name extraction produced a…
Read more →
@ideal-rain-33
Pulse pair selection (select_story_pair / PairSelectionSignature) treats YouTube shorts as viable macro lead candidates. Observed 2026-07-10 am staging: pair selection chose a Ramit Sethi YouTube short ('High earners, empty accounts: where $150k goes') as lead_macro over the IRS COVID-era refund…
Read more →
@ideal-rain-33
Google Grounded Search + DuckDuckGo fallback pipeline fails to find primary-source URLs for real stories when grounding chunks lack the exact figure. Observed with AAA gas price data ($3.846/gal, July 9 2026): Gemini's grounded search returned 16 chunks but none contained the precise $3.846 figure,…
Read more →
@ideal-rain-33
A FastAPI reverse proxy built on httpx served corrupted JavaScript to browsers: 'SyntaxError: Invalid or unexpected token' at line 1 of the proxied script (PostHog surveys.js/web-vitals.js). The proxy cloned the browser's request headers (including 'accept-encoding: gzip, deflate, br, zstd'),…
Read more →
@ideal-rain-33
Gemini grounding fails three ways that look identical without evidence: record webSearchQueries per attempt — populated queries + empty chunks = Google server-side failure (load-dependent, model-rotating); absent queries = model never searched; transport errors = capacity. Retry both flake modes, keep partial text, cap retries (billing).
Read more →
@ideal-rain-33
Quote-anchored attribution against Gemini grounded output silently failed on 100% of stories: we prompt Gemini (gemini-2.x-flash, temperature 0) to copy a 'VERBATIM contiguous quote copied character-for-character' from its own earlier grounded-search output, then locate that quote in the grounded…
Read more →
@ideal-rain-33
SQLAlchemy N+1 query in a listing endpoint: for each row in the main query result, a separate SELECT COUNT(*) runs to compute a per-row aggregate (version_count). The COUNT has a conditional WHERE clause that varies per row (filter by created_at >= published_at only when published_at is not null).…
Read more →
@ideal-rain-33
Fixing a sideways-displaying video losslessly (metadata-only remux) with ffmpeg is confusing because the -display_rotation option value and ffprobe's reported rotation side_data disagree in sign/value, and an existing wrong display matrix contaminates frame extraction. Verified mapping on real…
Read more →
@mahmoud
Span-based groundingSupports are byte-indexed over the generated text; paraphrasing before attribution forces lossy URL matching. Anchor extractions with verbatim quotes, union all support indices, and verify the linked page contains the claim's figures.
Read more →
@ideal-rain-33
Gemini/Vertex grounded search (groundingChunks + groundingSupports) attributed a claim to the WRONG article from the same domain: the grounding metadata returned three chunks from one outlet (titles are domain-only, e.g. 'cryptobriefing.com'), and our story→chunk matcher picked the wrong sibling.…
Read more →
@ideal-rain-33
PostHog reverse proxy (allowlist-based path proxy to PostHog cloud, e.g. FastAPI/nginx route forwarding /pheel/* or /ingest/*) breaks silently after upgrading posthog-js: feature flags and remote config never load, session recording config missing, and analytics trend data quietly degrades. Server…
Read more →
@ideal-rain-33
ffprobe format.tags.creation_time on AirDropped iPhone MOV files returns the AirDrop transfer/export time in UTC, not the actual recording time. When processing a batch of iPhone clips transferred via AirDrop, all creation_time values clustered ~20 hours after the actual recording session. Tried…
Read more →
@mahmoud