Spaces:

dhammadassana
/

tipitaka

Running

App Files Files Community

tipitaka / webapp /tipitaka-api

Commit History

May 31

07:59
Implement expandable volume breakdown with page chips in search panel Running

dhammawatthumpra

c1f9028
07:29
Fix Volume Breakdown query expansion discrepancy in SearchService

dhammawatthumpra

8c12cef
07:26
Implement hot-reloaded search rules config file for monk prefixes

dhammawatthumpra

7fd2d51
02:55
Fix synonym highlighting in snippet and reader panel, and fix health diagnostics 500 error

dhammawatthumpra

f33f622
00:08
Add health diag endpoint for troubleshooting vector search

dhammawatthumpra

aeed839

May 30

22:14
Implement synonym and related words highlighting in search results and reader panel

dhammawatthumpra

12b4775
22:11
Fix write permissions crash on Hugging Face Spaces by configuring paths before downloading and adding a writability check on /data

dhammawatthumpra

e99cf6e
21:33
Fix AI search by appending query expansions instead of replacing the original query

dhammawatthumpra

f955da2
21:17
fix: pass task='retrieval.query' to SentenceTransformers encode when using Jina or fine-tuned embedding models to prevent Task missing error

dhammawatthumpra

308d07e
21:15
fix: add peft to requirements.txt to support loading fine-tuned LoRA models in SentenceTransformers

dhammawatthumpra

4ffb8ff
20:36
feat: support Hugging Face persistent storage at /data and local Docker volumes for persisting SQLite database (search logs/cache) across container rebuilds

dhammawatthumpra

3dc49f8
19:56
feat: FTS5-first + AI button search with split button mode selector⁠

dhammawatthumpra

5e37f3b

May 27

13:05
fix: update QWEN_MODEL_ID default to 1.5B in main.py (missed in previous commit)

dhammawatthumpra

d5c52d4
10:44
chore: upgrade Qwen model default from 0.5B to 1.5B Instruct

dhammawatthumpra

ff9cfda

May 26

13:40
feat: complete turbovec search backend integration, ignore tvim file in git, and stabilize mobile UI transitions

dhammawatthumpra

603ead8
12:43
feat: add Google-style breadcrumb navigation to search results⁠

dhammawatthumpra

5c4239f
09:37
fix: resolve critical vector embedding bug and refine query transform prompt for prefix word matching

dhammawatthumpra

b00bdb2
09:08
fix: serialize concurrent thread access to in-memory SQLite connection to prevent race conditions and deadlocks

dhammawatthumpra

de8a899
08:29
Implement SQLite FTS5 prefix matching to support partial Thai word search and resolve segmentation tokenization issues

dhammawatthumpra

4c554ad
08:16
Swap default Query-Transform LLM to Qwen2.5-0.5B-Instruct to significantly reduce memory footprint and CPU generation latency on Hugging Face free tier

dhammawatthumpra

82e5f43
05:21
Optimize PyTorch resource usage: set CPU threads to 1 to prevent health check timeouts and reduce max generation tokens to 45

dhammawatthumpra

68c0ac1
05:10
Optimize backend preloading: preload Qwen & Jina-v5 at startup, cache Qwen in QueryTransformService class attributes, and bypass Ollama connection attempts when not ready

dhammawatthumpra

029c891
02:25
fix: optimize local mode Qdrant restore by replacing server snapshots with pre-packaged SQLite databases

dhammawatthumpra

cdcdbbf
00:54
docs: add SEARCH_ARCHITECTURE_V3 spec and include helper scripts for reindexing and RAG query testing

dhammawatthumpra

7805140

May 25

23:57
fix: increase Ollama HTTP client timeout to 30.0s to handle model cold starts resiliently

dhammawatthumpra

f1d32da
23:54
fix: implement query_batch_points and sequential fallback in SearchService, and raise 404 on page not found in pages router

dhammawatthumpra

3670990
22:45
feat: implement Search Architecture V3 (Hybrid search with Qwen query transformation, batch embeddings, batch Qdrant searches, and optimized CPU reranking)

dhammawatthumpra

c6f014f

May 19

08:26
fix: add Cache-Control headers to static files — no-cache for root icons, immutable for hashed assets

dhammawatthumpra

ed6d275
06:59
fix: serve all root-level static files (icons, favicon.svg, apple-touch-icons)

dhammawatthumpra

faab8c8

May 17

07:26
perf: skip LLM/RAG imports in abbrev lookup, query DB directly⁠

dhammawatthumpra

b8531b2
07:25
fix: use _default marker_id for non-numeric abbrev ids, skip RAG init⁠

dhammawatthumpra

f417652
07:19
refactor: disable LLM fallback for (ย่อ), fix JSON string array parsing⁠

dhammawatthumpra

efc32dd

May 16

22:35
fix: add อนุ. วิ. ปฏิ. dialogue markers to split formatting⁠

dhammawatthumpra

9babb2f
22:30
fix: exclude '.' from marker normalization preceding char⁠

dhammawatthumpra

70f46ee
21:58
fix: merge adjacent <B> headings with <I> footnote refs, fix marker normalization exclusion list⁠

dhammawatthumpra

6b67cae
08:23
feat: normalize reference extraction and popup UI

dhammawatthumpra

165565d
07:25
feat: enhance footnote extraction and display at bottom of page

dhammawatthumpra

23b4d13
07:19
feat: implement Reference Lookup API and ReferencePopup integration

dhammawatthumpra

b1f23a6
07:07
feat: migrate jina v5, refine AI popup, add reference popup UI

dhammawatthumpra

d2c65ec

May 15

02:45
feat: add sentence-transformers fallback for embedding when Ollama unavailable

dhammawatthumpra

abb8b56

May 7

19:20
fix: cleanup stale Qdrant lock on startup to prevent health alert

dhammawatthumpra

e04d706
09:17
fix: extract end markers from raw <B> tags instead of sections (works per-page, not proximity guess)

dhammawatthumpra

bf53ae4
09:15
fix: end markers only show for nearby sections (page proximity filter)

dhammawatthumpra

cfd4523
08:42
feat: first-page reader layout (centered homage, split volume title) + clean section range markers

dhammawatthumpra

ef0ec43
05:57
feat: use contents table for reader section hierarchy (L0+L1)⁠

dhammawatthumpra

575f508
04:46
feat: use contents table for full section hierarchy in reader (sections array instead of single title)

dhammawatthumpra

c9166a1
03:39
feat: markdown rendering for AI responses

dhammawatthumpra

0f2c05e
01:14
refactor: clean legacy, sync deps, add embedding cache, split RAGService, split search_service, add tests⁠

dhammawatthumpra

cb7868a

May 6

21:06
Refactor: Make app settings cloud-compatible by removing hardcoded Windows paths

dhammawatthumpra

f50c177
20:40
Refactor: Remove remaining hardcoded Windows paths and improve settings flexibility in RAGService

dhammawatthumpra

0860ede

Commit History

May 31

Implement expandable volume breakdown with page chips in search panel Running

Fix Volume Breakdown query expansion discrepancy in SearchService

Implement hot-reloaded search rules config file for monk prefixes

Fix synonym highlighting in snippet and reader panel, and fix health diagnostics 500 error

Add health diag endpoint for troubleshooting vector search

May 30

Implement synonym and related words highlighting in search results and reader panel

Fix write permissions crash on Hugging Face Spaces by configuring paths before downloading and adding a writability check on /data

Fix AI search by appending query expansions instead of replacing the original query

fix: pass task='retrieval.query' to SentenceTransformers encode when using Jina or fine-tuned embedding models to prevent Task missing error

fix: add peft to requirements.txt to support loading fine-tuned LoRA models in SentenceTransformers

feat: support Hugging Face persistent storage at /data and local Docker volumes for persisting SQLite database (search logs/cache) across container rebuilds

feat: FTS5-first + AI button search with split button mode selector⁠

May 27

fix: update QWEN_MODEL_ID default to 1.5B in main.py (missed in previous commit)

chore: upgrade Qwen model default from 0.5B to 1.5B Instruct

May 26

feat: complete turbovec search backend integration, ignore tvim file in git, and stabilize mobile UI transitions

feat: add Google-style breadcrumb navigation to search results⁠

fix: resolve critical vector embedding bug and refine query transform prompt for prefix word matching

fix: serialize concurrent thread access to in-memory SQLite connection to prevent race conditions and deadlocks

Implement SQLite FTS5 prefix matching to support partial Thai word search and resolve segmentation tokenization issues

Swap default Query-Transform LLM to Qwen2.5-0.5B-Instruct to significantly reduce memory footprint and CPU generation latency on Hugging Face free tier

Optimize PyTorch resource usage: set CPU threads to 1 to prevent health check timeouts and reduce max generation tokens to 45

Optimize backend preloading: preload Qwen & Jina-v5 at startup, cache Qwen in QueryTransformService class attributes, and bypass Ollama connection attempts when not ready

fix: optimize local mode Qdrant restore by replacing server snapshots with pre-packaged SQLite databases

docs: add SEARCH_ARCHITECTURE_V3 spec and include helper scripts for reindexing and RAG query testing

May 25

fix: increase Ollama HTTP client timeout to 30.0s to handle model cold starts resiliently

fix: implement query_batch_points and sequential fallback in SearchService, and raise 404 on page not found in pages router

feat: implement Search Architecture V3 (Hybrid search with Qwen query transformation, batch embeddings, batch Qdrant searches, and optimized CPU reranking)

May 19

fix: add Cache-Control headers to static files — no-cache for root icons, immutable for hashed assets

fix: serve all root-level static files (icons, favicon.svg, apple-touch-icons)

May 17

perf: skip LLM/RAG imports in abbrev lookup, query DB directly⁠

fix: use _default marker_id for non-numeric abbrev ids, skip RAG init⁠

refactor: disable LLM fallback for (ย่อ), fix JSON string array parsing⁠

May 16

fix: add อนุ. วิ. ปฏิ. dialogue markers to split formatting⁠

fix: exclude '.' from marker normalization preceding char⁠

fix: merge adjacent <B> headings with <I> footnote refs, fix marker normalization exclusion list⁠

feat: normalize reference extraction and popup UI

feat: enhance footnote extraction and display at bottom of page

feat: implement Reference Lookup API and ReferencePopup integration

feat: migrate jina v5, refine AI popup, add reference popup UI

May 15

feat: add sentence-transformers fallback for embedding when Ollama unavailable

May 7

fix: cleanup stale Qdrant lock on startup to prevent health alert

fix: extract end markers from raw <B> tags instead of sections (works per-page, not proximity guess)

fix: end markers only show for nearby sections (page proximity filter)

feat: first-page reader layout (centered homage, split volume title) + clean section range markers

feat: use contents table for reader section hierarchy (L0+L1)⁠

feat: use contents table for full section hierarchy in reader (sections array instead of single title)

feat: markdown rendering for AI responses

refactor: clean legacy, sync deps, add embedding cache, split RAGService, split search_service, add tests⁠

May 6

Refactor: Make app settings cloud-compatible by removing hardcoded Windows paths

Refactor: Remove remaining hardcoded Windows paths and improve settings flexibility in RAGService