simple-dynamsoft-mcp 5.0.0 → 6.3.0

package/.env.example

@@ -1,57 +1,89 @@
-# RAG provider selection (primary + fallback)
-# RAG_PROVIDER: auto | gemini | local | fuse
-#   - auto: gemini if GEMINI_API_KEY is set, otherwise local
-#   - gemini: remote embeddings via Google Gemini
-#   - local: local embeddings via @xenova/transformers
-#   - fuse: legacy fuzzy search only (no embeddings)
-RAG_PROVIDER=auto
-
-# RAG_FALLBACK: none | fuse | local
-#   - none: do not fall back if primary fails
-#   - fuse: fuzzy search fallback (fast, no embeddings)
-#   - local: local embeddings fallback (if primary is gemini)
-RAG_FALLBACK=fuse
-
-# Gemini remote embeddings (required when RAG_PROVIDER=gemini)
-# * GEMINI_API_KEY: your Gemini API key (never commit to git)
-# * GEMINI_EMBED_MODEL: models/embedding-001 | models/gemini-embedding-001 | gemini-embedding-001
-# * GEMINI_API_BASE_URL: override for proxies (default https://generativelanguage.googleapis.com)
-# * GEMINI_EMBED_BATCH_SIZE: batch size for Gemini embedding requests
-# GEMINI_API_KEY=your_key
-# GEMINI_EMBED_MODEL=gemini-embedding-001
-# GEMINI_API_BASE_URL=https://generativelanguage.googleapis.com
-# GEMINI_EMBED_BATCH_SIZE=16
-
-# Local embeddings (used when RAG_PROVIDER=local or fallback=local)
-# * RAG_LOCAL_MODEL: Hugging Face model id (default Xenova/all-MiniLM-L6-v2)
-# * RAG_LOCAL_QUANTIZED: true|false (smaller/faster model download when true)
-# RAG_LOCAL_MODEL=Xenova/all-MiniLM-L6-v2
-# RAG_LOCAL_QUANTIZED=true
-
-# Cache locations
-# * RAG_CACHE_DIR: vector index cache (default data/.rag-cache)
-# * RAG_MODEL_CACHE_DIR: local model cache (default data/.rag-cache/models)
-# RAG_CACHE_DIR=data/.rag-cache
-# RAG_MODEL_CACHE_DIR=data/.rag-cache/models
-
-# Indexing + retrieval tuning
-# * RAG_CHUNK_SIZE: max chars per chunk when embedding doc content
-# * RAG_CHUNK_OVERLAP: overlap between chunks (helps context continuity)
-# * RAG_MAX_CHUNKS_PER_DOC: cap chunks per doc to control index size
-# * RAG_MAX_TEXT_CHARS: max chars per embedding input
-# * RAG_MIN_SCORE: minimum cosine similarity to keep a hit (0 disables filtering). Default 0.2.
-# * RAG_INCLUDE_SCORE: include similarity score in search results (debugging)
-# RAG_CHUNK_SIZE=1200
-# RAG_CHUNK_OVERLAP=200
-# RAG_MAX_CHUNKS_PER_DOC=6
-# RAG_MAX_TEXT_CHARS=4000
-# RAG_MIN_SCORE=0.2
-# RAG_INCLUDE_SCORE=false
-
-# Cache/boot controls
-# * RAG_REBUILD: true to ignore cache and rebuild on startup/search
-# * RAG_PREWARM: true to build the embedding index at startup
-# * RAG_PREWARM_BLOCK: true to block startup until prewarm completes
-# RAG_REBUILD=false
-# RAG_PREWARM=false
-# RAG_PREWARM_BLOCK=false
+# RAG provider selection (primary + fallback)
+# RAG_PROVIDER: auto | gemini | local | fuse
+#   - auto: gemini if GEMINI_API_KEY is set, otherwise local
+#   - gemini: remote embeddings via Google Gemini
+#   - local: local embeddings via @xenova/transformers
+#   - fuse: legacy fuzzy search only (no embeddings)
+RAG_PROVIDER=auto
+
+# RAG_FALLBACK: none | fuse | local
+#   - none: do not fall back if primary fails
+#   - fuse: fuzzy search fallback (fast, no embeddings)
+#   - local: local embeddings fallback (if primary is gemini)
+RAG_FALLBACK=fuse
+
+# Gemini remote embeddings (required when RAG_PROVIDER=gemini)
+# * GEMINI_API_KEY: your Gemini API key (never commit to git)
+# * GEMINI_EMBED_MODEL: models/embedding-001 | models/gemini-embedding-001 | gemini-embedding-001
+# * GEMINI_API_BASE_URL: override for proxies (default https://generativelanguage.googleapis.com)
+# * GEMINI_EMBED_BATCH_SIZE: batch size for Gemini embedding requests
+# * GEMINI_RETRY_MAX_ATTEMPTS: max retry attempts for retryable Gemini errors
+# * GEMINI_RETRY_BASE_DELAY_MS: base delay for exponential backoff
+# * GEMINI_RETRY_MAX_DELAY_MS: max delay cap for exponential backoff
+# * GEMINI_REQUEST_THROTTLE_MS: fixed delay between Gemini requests
+# GEMINI_API_KEY=your_key
+# GEMINI_EMBED_MODEL=gemini-embedding-001
+# GEMINI_API_BASE_URL=https://generativelanguage.googleapis.com
+# GEMINI_EMBED_BATCH_SIZE=16
+# GEMINI_RETRY_MAX_ATTEMPTS=5
+# GEMINI_RETRY_BASE_DELAY_MS=500
+# GEMINI_RETRY_MAX_DELAY_MS=10000
+# GEMINI_REQUEST_THROTTLE_MS=0
+
+# Local embeddings (used when RAG_PROVIDER=local or fallback=local)
+# * RAG_LOCAL_MODEL: Hugging Face model id (default Xenova/all-MiniLM-L6-v2)
+# * RAG_LOCAL_QUANTIZED: true|false (smaller/faster model download when true)
+# RAG_LOCAL_MODEL=Xenova/all-MiniLM-L6-v2
+# RAG_LOCAL_QUANTIZED=true
+
+# Cache locations
+# * RAG_CACHE_DIR: vector index cache (default data/.rag-cache)
+# * RAG_MODEL_CACHE_DIR: local model cache (default data/.rag-cache/models)
+# RAG_CACHE_DIR=data/.rag-cache
+# RAG_MODEL_CACHE_DIR=data/.rag-cache/models
+
+# Indexing + retrieval tuning
+# * RAG_CHUNK_SIZE: max chars per chunk when embedding doc content
+# * RAG_CHUNK_OVERLAP: overlap between chunks (helps context continuity)
+# * RAG_MAX_CHUNKS_PER_DOC: cap chunks per doc to control index size
+# * RAG_MAX_TEXT_CHARS: max chars per embedding input
+# * RAG_MIN_SCORE: minimum cosine similarity to keep a hit (0 disables filtering). Default 0.2.
+# * RAG_INCLUDE_SCORE: include similarity score in search results (debugging)
+# RAG_CHUNK_SIZE=1200
+# RAG_CHUNK_OVERLAP=200
+# RAG_MAX_CHUNKS_PER_DOC=6
+# RAG_MAX_TEXT_CHARS=4000
+# RAG_MIN_SCORE=0.2
+# RAG_INCLUDE_SCORE=false
+
+# Cache/boot controls
+# * RAG_REBUILD: true to ignore cache and rebuild on startup/search
+# * RAG_PREWARM: true to build the embedding index at startup
+# * RAG_PREWARM_BLOCK: true to block startup until prewarm completes
+# * RAG_PREBUILT_INDEX_AUTO_DOWNLOAD: auto-download prebuilt local index when local embeddings are selected
+# * RAG_PREBUILT_INDEX_URL: override prebuilt index archive URL (default GitHub release asset for current package version)
+# * RAG_PREBUILT_INDEX_TIMEOUT_MS: timeout for prebuilt index download request
+# RAG_REBUILD=false
+# RAG_PREWARM=false
+# RAG_PREWARM_BLOCK=false
+# RAG_PREBUILT_INDEX_AUTO_DOWNLOAD=true
+# RAG_PREBUILT_INDEX_URL=
+# RAG_PREBUILT_INDEX_TIMEOUT_MS=180000
+
+# Optional data submodule sync on server startup
+# * DATA_SYNC_ON_START: true to fetch + fast-forward configured submodules
+# * DATA_SYNC_TIMEOUT_MS: timeout per git operation in milliseconds
+# DATA_SYNC_ON_START=false
+# DATA_SYNC_TIMEOUT_MS=30000
+
+# Data bootstrap mode (for npm/npx installs without submodules)
+# * MCP_DATA_DIR: explicit existing data root (must contain metadata/, samples/, documentation/)
+# * MCP_DATA_AUTO_DOWNLOAD: auto-download pinned sample/doc archives when local data is missing
+# * MCP_DATA_CACHE_DIR: where downloaded data is stored
+# * MCP_DATA_REFRESH_ON_START: force re-download even when cache matches lock manifest
+# * MCP_DATA_DOWNLOAD_TIMEOUT_MS: timeout per archive download request in milliseconds
+# MCP_DATA_DIR=
+# MCP_DATA_AUTO_DOWNLOAD=true
+# MCP_DATA_CACHE_DIR=
+# MCP_DATA_REFRESH_ON_START=false
+# MCP_DATA_DOWNLOAD_TIMEOUT_MS=180000

package/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2026 Xiao Ling
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2026 Xiao Ling
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.