Merge pull request #30 from GaiaNet-AI/multiple_qdrant_servers

GaiaNet-AI · Apr 13, 2024 · 61e9036 · 61e9036
2 parents 6a39cf9 + 1900b1e
commit 61e9036
Show file tree

Hide file tree

Showing 4 changed files with 80 additions and 46 deletions.
diff --git a/config.json b/config.json
@@ -9,6 +9,7 @@
     "embedding": "https://huggingface.co/second-state/All-MiniLM-L6-v2-Embedding-GGUF/resolve/main/all-MiniLM-L6-v2-ggml-model-f16.gguf",
     "embedding_ctx_size": "384",
     "snapshot": "https://huggingface.co/datasets/gaianet/paris/resolve/main/paris_384_all-minilm-l6-v2_f16.snapshot",
+    "embedding_collection_name": "default",
     "rag_prompt": "Use the following pieces of context to answer the user's question.\nIf you don't know the answer, just say that you don't know, don't try to make up an answer.\n----------------\n",
     "domain": "gaianet.xyz",
     "llamaedge_port": "8080"

diff --git a/install.sh b/install.sh
@@ -10,12 +10,15 @@ cwd=$(pwd)
 reinstall=0
 # url to the config file
 config_url=""
+# path to the gaianet base directory
+gaianet_base_dir="$HOME/gaianet"
 
 function print_usage {
     printf "Usage:\n"
     printf "  ./install.sh [Options]\n\n"
     printf "Options:\n"
     printf "  --config <Url>: specify a url to the config file\n"
+    printf "  --base <Path>: specify a path to the gaianet base directory\n"
     printf "  --reinstall: install and download all required deps\n"
     printf "  --help: Print usage\n"
 }
@@ -28,6 +31,11 @@ while [[ $# -gt 0 ]]; do
             shift
             shift
             ;;
+        --base)
+            gaianet_base_dir="$2"
+            shift
+            shift
+            ;;
         --reinstall)
             reinstall=1
             shift
@@ -44,12 +52,8 @@ while [[ $# -gt 0 ]]; do
     esac
 done
 
-
 printf "\n"
 
-# Set "gaianet_base_dir" to $HOME/gaianet
-gaianet_base_dir="$HOME/gaianet"
-
 # if need to reinstall, remove the $gaianet_base_dir directory
 if [ "$reinstall" -eq 1 ] && [ -d "$gaianet_base_dir" ]; then
     printf "[+] Removing the existing $gaianet_base_dir directory ...\n\n"
@@ -240,11 +244,13 @@ fi
 # 10. recover from the given qdrant collection snapshot =======================
 printf "[+] Initializing the Qdrant server ...\n\n"
 
-# check 6333 port is in use or not
+qdrant_pid=0
+qdrant_already_running=false
 if [ "$(uname)" == "Darwin" ] || [ "$(expr substr $(uname -s) 1 5)" == "Linux" ]; then
     if lsof -Pi :6333 -sTCP:LISTEN -t >/dev/null ; then
-        printf "It appears that the GaiaNet node is running. Please stop it first.\n\n"
-        exit 1
+        # printf "It appears that the GaiaNet node is running. Please stop it first.\n\n"
+        # exit 1
+	qdrant_already_running=true
     fi
 elif [ "$(expr substr $(uname -s) 1 10)" == "MINGW32_NT" ]; then
     printf "For Windows users, please run this script in WSL.\n"
@@ -254,15 +260,18 @@ else
     exit 1
 fi
 
-# start qdrant
-cd $gaianet_base_dir/qdrant
-nohup $gaianet_base_dir/bin/qdrant > $log_dir/init-qdrant.log 2>&1 &
-sleep 15
-qdrant_pid=$!
+if [ "$qdrant_already_running" = false ]; then
+    # start qdrant
+    cd $gaianet_base_dir/qdrant
+    nohup $gaianet_base_dir/bin/qdrant > $log_dir/init-qdrant.log 2>&1 &
+    sleep 15
+    qdrant_pid=$!
+fi
 
 cd $gaianet_base_dir
 url_snapshot=$(awk -F'"' '/"snapshot":/ {print $4}' config.json)
 url_document=$(awk -F'"' '/"document":/ {print $4}' config.json)
+embedding_collection_name=$(awk -F'"' '/"embedding_collection_name":/ {print $4}' config.json)
 
 if [ -n "$url_snapshot" ]; then
     # 10.1 recover from the given qdrant collection snapshot
@@ -271,18 +280,18 @@ if [ -n "$url_snapshot" ]; then
     curl --progress-bar -L $url_snapshot -o default.snapshot
 
     cd $gaianet_base_dir
-    # remove the 'default' collection if it exists
-    del_response=$(curl -s -X DELETE http://localhost:6333/collections/default \
+    # remove the collection if it exists
+    del_response=$(curl -s -X DELETE http://localhost:6333/collections/$embedding_collection_name \
         -H "Content-Type: application/json")
     status=$(echo "$del_response" | grep -o '"status":"[^"]*"' | cut -d':' -f2 | tr -d '"')
     if [ "$status" != "ok" ]; then
-        printf "    Failed to remove the 'default' collection. $del_response\n\n"
+        printf "    Failed to remove the $embedding_collection_name collection. $del_response\n\n"
         kill $qdrant_pid
         exit 1
     fi
 
     # Import the default.snapshot file
-    response=$(curl -s -X POST 'http://localhost:6333/collections/default/snapshots/upload?priority=snapshot' \
+    response=$(curl -s -X POST http://localhost:6333/collections/$embedding_collection_name/snapshots/upload?priority=snapshot \
         -H 'Content-Type:multipart/form-data' \
         -F 'snapshot=@default.snapshot')
     sleep 5
@@ -301,14 +310,14 @@ elif [ -n "$url_document" ]; then
 
     printf "[+] Creating a Qdrant collection from the given document ...\n\n"
 
-    # Remove the 'default' collection if it exists
-    printf "    * Removing 'default' collection if it exists ...\n\n"
+    # Remove the collection if it exists
+    printf "    * Removing collection if it exists ...\n\n"
     # remove the 'default' collection if it exists
-    del_response=$(curl -s -X DELETE http://localhost:6333/collections/default \
+    del_response=$(curl -s -X DELETE http://localhost:6333/collections/$embedding_collection_name \
         -H "Content-Type: application/json")
     status=$(echo "$del_response" | grep -o '"status":"[^"]*"' | cut -d':' -f2 | tr -d '"')
     if [ "$status" != "ok" ]; then
-        printf "    Failed to remove the 'default' collection. $del_response\n\n"
+        printf "    Failed to remove the collection. $del_response\n\n"
         kill $qdrant_pid
         exit 1
     fi
@@ -337,6 +346,8 @@ elif [ -n "$url_document" ]; then
     embedding_model_stem=$(basename "$embedding_model_name" .gguf)
     # parse context size for embedding model
     embedding_ctx_size=$(awk -F'"' '/"embedding_ctx_size":/ {print $4}' config.json)
+    # parse cli options for embedding vector collection name
+    embedding_collection_name=$(awk -F'"' '/"embedding_collection_name":/ {print $4}' config.json)
     # parse port for LlamaEdge API Server
     llamaedge_port=$(awk -F'"' '/"llamaedge_port":/ {print $4}' config.json)
 
@@ -370,6 +381,7 @@ elif [ -n "$url_document" ]; then
     rag-api-server.wasm -p $prompt_type \
     --model-name $chat_model_stem,$embedding_model_stem \
     --ctx-size $chat_ctx_size,$embedding_ctx_size \
+    --qdrant-collection-name $embedding_collection_name \
     --web-ui ./dashboard \
     --socket-addr 0.0.0.0:$llamaedge_port \
     --log-prompts \
@@ -424,8 +436,10 @@ else
 fi
 printf "\n"
 
-# stop qdrant
-kill $qdrant_pid
+if [ "$qdrant_already_running" = false ]; then
+    # stop qdrant
+    kill $qdrant_pid
+fi
 
 # ======================================================================================
 

diff --git a/start.sh b/start.sh
@@ -1,6 +1,9 @@
 #!/bin/bash
 
+# default is to use frpc
 local_only=0
+# path to the gaianet base directory
+gaianet_base_dir="$HOME/gaianet"
 
 while [[ $# -gt 0 ]]; do
     key="$1"
@@ -9,6 +12,11 @@ while [[ $# -gt 0 ]]; do
             local_only=1
             shift
             ;;
+	--base)
+            gaianet_base_dir="$2"
+            shift
+            shift
+            ;;
         *)
             echo "Unknown argument: $key"
             print_usage
@@ -21,13 +29,11 @@ done
 # represents the directory where the script is located
 script_dir=$(pwd)
 
-# Check if "gaianet" directory exists in $HOME
-if [ ! -d "$HOME/gaianet" ]; then
-    printf "Not found $HOME/gaianet\n"
+# Check if "gaianet" home directory exists
+if [ ! -d "$gaianet_base_dir" ]; then
+    printf "Not found $gaianet_base_dir\n"
     exit 1
 fi
-# Set "gaianet_base_dir" to $HOME/gaianet
-gaianet_base_dir="$HOME/gaianet"
 
 # check if `log` directory exists or not
 if [ ! -d "$gaianet_base_dir/log" ]; then
@@ -44,11 +50,13 @@ fi
 # 1. start a Qdrant instance
 printf "[+] Starting Qdrant instance ...\n"
 
+qdrant_already_running=false
 if [ "$(uname)" == "Darwin" ] || [ "$(expr substr $(uname -s) 1 5)" == "Linux" ]; then
     if lsof -Pi :6333 -sTCP:LISTEN -t >/dev/null ; then
-        printf "    Port 6333 is in use. Stopping the process on 6333 ...\n\n"
-        pid=$(lsof -t -i:6333)
-        kill -9 $pid
+        # printf "    Port 6333 is in use. Stopping the process on 6333 ...\n\n"
+        # pid=$(lsof -t -i:6333)
+        # kill -9 $pid
+        qdrant_already_running=true
     fi
 elif [ "$(expr substr $(uname -s) 1 10)" == "MINGW32_NT" ]; then
     printf "For Windows users, please run this script in WSL.\n"
@@ -58,17 +66,19 @@ else
     exit 1
 fi
 
-qdrant_executable="$gaianet_base_dir/bin/qdrant"
-if [ -f "$qdrant_executable" ]; then
-    cd $gaianet_base_dir/qdrant
-    nohup $qdrant_executable > $log_dir/start-qdrant.log 2>&1 &
-    sleep 2
-    qdrant_pid=$!
-    echo $qdrant_pid > $gaianet_base_dir/qdrant.pid
-    printf "\n    Qdrant instance started with pid: $qdrant_pid\n\n"
-else
-    printf "Qdrant binary not found at $qdrant_executable\n\n"
-    exit 1
+if [ "$qdrant_already_running" = false ]; then
+    qdrant_executable="$gaianet_base_dir/bin/qdrant"
+    if [ -f "$qdrant_executable" ]; then
+        cd $gaianet_base_dir/qdrant
+        nohup $qdrant_executable > $log_dir/start-qdrant.log 2>&1 &
+        sleep 2
+        qdrant_pid=$!
+        echo $qdrant_pid > $gaianet_base_dir/qdrant.pid
+        printf "\n    Qdrant instance started with pid: $qdrant_pid\n\n"
+    else
+        printf "Qdrant binary not found at $qdrant_executable\n\n"
+        exit 1
+    fi
 fi
 
 # 2. start a LlamaEdge instance
@@ -94,6 +104,8 @@ rag_prompt=$(awk -F'"' '/"rag_prompt":/ {print $4}' config.json)
 reverse_prompt=$(awk -F'"' '/"reverse_prompt":/ {print $4}' config.json)
 # parse cli options for embedding model
 url_embedding_model=$(awk -F'"' '/"embedding":/ {print $4}' config.json)
+# parse cli options for embedding vector collection name
+embedding_collection_name=$(awk -F'"' '/"embedding_collection_name":/ {print $4}' config.json)
 # gguf filename
 embedding_model_name=$(basename $url_embedding_model)
 # stem part of the filename
@@ -133,6 +145,7 @@ cmd=(wasmedge --dir .:./dashboard \
   --model-name $chat_model_stem,$embedding_model_stem \
   --ctx-size $chat_ctx_size,$embedding_ctx_size \
   --prompt-template $prompt_type \
+  --qdrant-collection-name $embedding_collection_name \
   --web-ui ./ \
   --socket-addr 0.0.0.0:$llamaedge_port \
   --log-prompts \

diff --git a/stop.sh b/stop.sh
@@ -2,11 +2,14 @@
 
 # 1: stop WasmEdge, Qdrant and frpc processes
 force_stop=0
+# path to the gaianet base directory
+gaianet_base_dir="$HOME/gaianet"
 
 function print_usage {
     printf "Usage:\n"
     printf "  ./stop.sh [--force-stop]\n\n"
     printf "  --force:  stop WasmEdge, Qdrant and frpc processes\n"
+    printf "  --base <Path>:  stop GaiaNet node installed in a specified home directory\n"
 }
 
 while [[ $# -gt 0 ]]; do
@@ -16,6 +19,11 @@ while [[ $# -gt 0 ]]; do
             force_stop=1
             shift
             ;;
+	--base)
+            gaianet_base_dir="$2"
+            shift
+            shift
+            ;;
         --help)
             print_usage
             exit 0
@@ -29,12 +37,10 @@ while [[ $# -gt 0 ]]; do
 done
 
 # Check if "gaianet" directory exists in $HOME
-if [ ! -d "$HOME/gaianet" ]; then
-    printf "Not found $HOME/gaianet\n"
+if [ ! -d "$gaianet_base_dir" ]; then
+    printf "Not found $gaianet_base_dir\n"
     exit 1
 fi
-# Set "gaianet_base_dir" to $HOME/gaianet
-gaianet_base_dir="$HOME/gaianet"
 
 if [ $force_stop -eq 1 ]; then
     printf "Force stopping WasmEdge, Qdrant and frpc processes ...\n"
@@ -84,4 +90,4 @@ else
     fi
 fi
 
-exit 0
+exit 0