zovos/vk_hackathon
1
0
Fork 2
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Compare commits

base: zovos:4800e25dd622ac9760d2b4b09c36cfe6c910ae0a
Branches Tags
zovos:main
zovos:best-0.5541
zovos:best-0.5496
zovos:v-rerank17
zovos:v-pushed-multifile
zovos:v-multifile-architecture
zovos:v-4core-optimized
zovos:v1.0-working
..
compare: zovos:8d9b0d5f6315cff34785d7a6252e07162499ef94
Branches Tags
zovos:main
zovos:best-0.5541
zovos:best-0.5496
zovos:v-rerank17
zovos:v-pushed-multifile
zovos:v-multifile-architecture
zovos:v-4core-optimized
zovos:v1.0-working

5 commits
4800e25dd6 ... 8d9b0d5f63

Author SHA1 Message Date
q
8d9b0d5f63 best: score 0.5541 (recall 0.5759, ndcg 0.4670) 
- DENSE_PREFETCH_K 80 → 120
- RERANK_LIMIT 25 → 35
- add question.text as extra dense when differs from search_text

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-19 13:23:03 +03:00
q
2a274e93d8 best: score 0.5496 (recall 0.5698, ndcg 0.4690) 
Search improvements on top of v1.0 index:
- RERANK_LIMIT 17 → 25
- prefilter with keyword-boosted stragglers (KEYWORD_BOOST_EXTRA=10)
- dense/sparse queries prefer search_text, keywords always in sparse
- variants+hyde as extra dense queries (up to 3)
- message_id score aggregation (rerank head full RRF, tail with k=60)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-19 12:42:29 +03:00
q
6869189099 Revert to v1.0-working + RERANK_LIMIT 15→17 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 23:02:05 +03:00
q
7c2fb0ccc4 Fix 500 error: replace datetime_range with range in Qdrant filter 
qdrant-client 1.15.1 does not support datetime_range in FieldCondition.
Use models.Range with string comparison (same as Lotus reference).
Also wrap date filter in try-except to prevent crash on bad date format.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 22:01:34 +03:00
q
e468c4768b Migrate to multi-file architecture: smarter chunking + fixed RERANK_LIMIT 
index: message-based windowed chunking (5 msgs/1h gap), better unicode
cleaning, separate dense (with timestamps)/sparse content renderers,
BM25 preload on startup, ThreadPoolExecutor(4), UVICORN_WORKERS=4,
Dockerfile copies all *.py

search: proper multi-module structure (query_builder, retrieval, rerank,
aggregation), RERANK_LIMIT 60→15 (fixes 429 errors), extra dense vectors
for variants/hyde, date+asker metadata filters, httpx pool (100/20/30s),
BM25 preload on startup, Dockerfile copies all *.py

68/68 unit tests passing

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 21:49:14 +03:00
1 changed files with 1 additions and 0 deletions
Show stats Expand all files Collapse all files

1
Lotus_it-academy-hackathon-solution Submodule

@ -0,0 +1 @@
Subproject commit 3b9ff54ccf1540f8259a438b52d5bb05553e9dcb