debug

Dockerfile

@@ -61,15 +61,19 @@ ENV PATH="/usr/bin/local:/app/text2vec-transformers:/app/text2vec-transformers/b
 #RUN pip install nltk==3.8.1 optimum==1.13.2 onnxruntime==1.16.1 onnx==1.14.1
 RUN ./custom_prerequisites.py
 
-
 ##############################
 RUN useradd -m -u 1000 user
 
+chmod -R 755  /app
+chown -R user /app
+chgrp -R user /app
+
 #############################################
 # Specify /data volume.
 #VOLUME /data
 
 WORKDIR /app
+USER user
 
 ##############################################################################
 # Start the weaviate vector database, text2vec-transformers and the semantic search app.

app.py

@@ -36,7 +36,7 @@ try:
 
 
     def runStartup():
-        logger.info("### Running startup.sh")
+        logger.debug("### Running startup.sh")
         result = ""
         try:
             #result = subprocess.run("/app/startup.sh",shell=False,capture_output=None,text=None,timeout=300)
@@ -55,7 +55,7 @@ try:
             except Exception as e2:
                 emsg = str(e2)
                 logger.error(f"#### Displaying startup.log EXCEPTION. e2: {emsg}")
-        logger.info("### Running startup.sh complete")
+        logger.debug("### Running startup.sh complete")
     if 'runStartup' not in st.session_state:
         st.session_state.runStartup = True
         runStartup()
@@ -69,8 +69,10 @@ try:
 
     # Function to load the CSS file
     def load_css(file_name):
+        logger.debug("#### load_css entered.")
         with open(file_name) as f:
             st.markdown(f'<style>{f.read()}</style>', unsafe_allow_html=True)
+        logger.debug("#### load_css exited.")
 
     # Load the custom CSS
     if 'load_css' not in st.session_state:
@@ -93,7 +95,7 @@ try:
     # Connect to the Weaviate vector database.
     #if 'client' not in st.session_state:
     if 'client' not in st.session_state:
-        logger.info("#### Create Weaviate db client connection.")
+        logger.debug("#### Create Weaviate db client connection.")
         client = weaviate.WeaviateClient(
             connection_params=ConnectionParams.from_params(
                 http_host="localhost",
@@ -109,6 +111,7 @@ try:
         )
         client.connect()
         st.session_state.client = client
+        logger.debug("#### Create Weaviate db client connection exited.")
     else:
         client = st.session_state.client
 
@@ -116,8 +119,8 @@ try:
     #######################################################
     # Read each text input file, parse it into a document,
     # chunk it, collect chunks and document name.
-    if not client.collections.exists("Documents") or not client.collections.exists("Documes") :
-        logger.info("#### Read and chunk input text files.")
+    if not client.collections.exists("Documents") or not client.collections.exists("Chunks) :
+        logger.debug("#### Read and chunk input text files.")
         for filename in os.listdir(pathString):
             logger.info(filename)
             path = Path(pathString + "/" + filename)
@@ -150,6 +153,7 @@ try:
 
             logger.debug(f"### filename, title: {filename}, {title}")
             logger.debug(f"### webpageDocNames: {webpageDocNames}")
+        logger.debug("#### Read and chunk input text files exited.")
 
 
 
@@ -158,7 +162,7 @@ try:
     #wpCollection = createWebpageCollection()
     #wpChunkCollection = createChunksCollection()
     if not client.collections.exists("Documents"):
-        logger.info("#### createWebpageCollection() entered.")
+        logger.debug("#### createWebpageCollection() entered.")
         #client.collections.delete("Documents")
         class_obj = {
           "class": "Documents",
@@ -207,10 +211,11 @@ try:
             ]  
         }
         wpCollection = client.collections.create_from_dict(class_obj)
+        logger.debug("#### createWebpageCollection() exited.")
 
 
     if not client.collections.exists("Chunks"):
-        logger.info("#### createChunksCollection() entered.")
+        logger.debug("#### createChunksCollection() entered.")
         #client.collections.delete("Chunks")
         class_obj = {
             "class": "Chunks",
@@ -258,12 +263,13 @@ try:
             ]
         }
         wpChunkCollection = client.collections.create_from_dict(class_obj)
+        logger.debug("#### createChunksCollection() exited.")
 
 
     ###########################################################
     # Create document and chunks objects in the database.
     if not client.collections.exists("Documents") :
-        logger.info("#### Create page/doc db objects.")
+        logger.debug("#### Create page/doc db objects.")
         for i, className in enumerate(webpageDocNames):
             title = webpageTitles[i]
             logger.debug(f"## className, title: {className}, {title}")
@@ -277,9 +283,10 @@ try:
                 "content": page_content
               }
             )
+        logger.debug("#### Create page/doc/db/objects exited.")
         
     if not client.collections.exists("Chunks") :
-        logger.info("#### Create chunk db objects.")
+        logger.debug("#### Create chunk db objects.")
         # Insert the chunks for the document.
         for i2, chunk in enumerate(webpageChunks[i]):
             chunk_uuid = wpChunkCollection.data.insert(
@@ -293,12 +300,14 @@ try:
                 }
               }
             )
+        logger.debug("#### Create chunk db objects exited.")
 
 
     #################################################################
     # Initialize the LLM.
     model_path = "/app/llama-2-7b-chat.Q4_0.gguf"
     if 'llm' not in st.session_state:
+        logger.debug("### Initializing LLM.")
         llm = Llama(model_path,
                     #*,
                     n_gpu_layers=0,
@@ -340,6 +349,7 @@ try:
                     verbose=True
                    )
         st.session_state.llm = llm
+        logger.debug("### Initializing LLM exited.")
     else:
         llm = st.session_state.llm
 
@@ -415,6 +425,7 @@ try:
             st.session_state.rspTA = st.text_area(label="rspTA",value=st.session_state.rspTA)
 
     def runLLM(prompt):
+        echo "### runLLM entered."
         max_tokens = 1000
         temperature = 0.3
         top_p = 0.1
@@ -430,6 +441,7 @@ try:
            stop=stop,
         )
         result = modelOutput["choices"][0]["text"].strip()
+        logger.debug("### runLLM exited.")
         return(result)
 
     def setPrompt(pprompt,ragFlag):
@@ -445,11 +457,13 @@ try:
         else:
             userPrompt = pprompt
         #prompt = f""" <s> [INST] <<SYS>> {systemTextArea.value} </SYS>> Q: {userPrompt} A: [/INST]"""
+        logger.debug("setPrompt exited.")
         return userPrompt
 
 
     def on_submitButton_clicked():
         logger = st.session_state.logger
+        logger.debug("### on_submitButton_clicked entered.")
         logger.debug("\n### on_submitButton_clicked")
         st.session_state.sysTAtext = st.session_state.sysTA
         logger.info(f"sysTAtext: {st.session_state.sysTAtext}")
@@ -461,13 +475,15 @@ try:
         st.session_state.rspTAtext = runLLM(st.session_state.userpTAtext)
         st.session_state.rspTA = st.session_state.rspTAtext
         logger.info(f"rspTAtext: {st.session_state.rspTAtext}")
+        
+        logger.debug("### on_submitButton_clicked exited.")
 
 
     with st.sidebar:
         st.selectRag    = st.checkbox("Enable Query With RAG",value=False,key="selectRag",help=None,on_change=None,args=None,kwargs=None,disabled=False,label_visibility="visible")
         st.submitButton = st.button("Run LLM Query",key=None,help=None,on_click=on_submitButton_clicked,args=None,kwargs=None,type="secondary",disabled=False,use_container_width=False)
 
-    logger.info("#### semsearch.py terminating.")
+    logger.info("#### semsearch.py end of code.")
 except Exception as e:
     try:
         emsg = str(e)