app factory and more

.gitignore

@@ -1,6 +1,10 @@
 backend/__pycache__/
 ollama
-deployment/ollama
 *.env
 deployment/*.env
+deployment/ollama
+# neo4j
+deployment/data
+deployment/logs
+
 

backend/lib/chatbot.py

@@ -87,26 +87,11 @@ def train_text(bot_id, text):
 
 
 
-#TODO add history
+
 
 def ask_bot(question, bot_id):
-    bot = Chatbot.get(id=bot_id)
-    llm = Ollama(
-        model=bot.llm_model,
-        base_url=ollama_url
-    )
-    query = bot.system_prompt + " " + question
-    for chunk in llm.stream(query):
-        yield chunk
-
-
-
-#connections.get_connection()
-#if es.indices.exists(index="index"):
-
-def ask_bot2(question, bot_id):
     """
-    Asks a chatbot
+    Asks a chatbot using RAG resources
     """
 
     bot = Chatbot.get(id=bot_id)
@@ -159,7 +144,6 @@ def ask_bot2(question, bot_id):
                 xs.append([score, dict(doc)])
             return xs
 
-
         return {
             "answer_generator": gen_func,
             "get_score_docs": get_score_docs

backend/lib/knowledge.py

@@ -0,0 +1,38 @@
+
+
+
+
+from neo4j import GraphDatabase
+
+"""
+pwd = "neo4j2"
+proto = "bolt"
+host = "192.168.99.101"
+
+driver = GraphDatabase.driver("%s://%s:7687" % (proto, host), auth=("neo4j", pwd), encrypted=False)
+
+def add_friend(tx, name, friend_name):
+    tx.run("MERGE (a:Person {name: $name}) "
+           "MERGE (a)-[:KNOWS]->(friend:Person {name: $friend_name})",
+           name=name, friend_name=friend_name)
+
+def print_friends(tx, name):
+    for record in tx.run("MATCH (a:Person)-[:KNOWS]->(friend) WHERE a.name = $name "
+                         "RETURN friend.name ORDER BY friend.name", name=name):
+        print(record["friend.name"])
+
+with driver.session() as session:
+    session.write_transaction(add_friend, "Arthur", "Guinevere")
+    session.write_transaction(add_friend, "Arthur", "Lancelot")
+    session.write_transaction(add_friend, "Arthur", "Merlin")
+    session.read_transaction(print_friends, "Arthur")
+
+driver.close()
+"""
+
+
+
+
+
+
+

backend/lib/logging.py

@@ -0,0 +1,68 @@
+import logging
+import os
+from datetime import datetime, date
+from lib.models import LogEntry
+
+
+
+
+class ElasticsearchLogHandler(logging.Handler):
+
+    def __init__(self, level):
+        logging.Handler.__init__(self=self)
+        #super().__init__(self=self)
+        self.setLevel(level)
+
+    def emit(self, record):
+
+        #print(str(record.__dict__), flush=True)
+
+        #{'name': 'werkzeug',
+        #  'msg': '192.168.64.1 - - [07/Sep/2024 11:43:23] "%s" %s %s', 
+        # 'args': ('GET /socket.io/?EIO=4&transport=websocket&sid=MtyTmZQs5IA6DnvhAAAA HTTP/1.1', '200', '-'), 
+        # 'levelname': 'INFO', 
+        # 'levelno': 20,
+        #  'pathname': '/usr/local/lib/python3.12/dist-packages/werkzeug/_internal.py', 
+        # 'filename': '_internal.py', 
+        # 'module': '_internal',
+        #  'exc_info': None, 
+        # 'exc_text': None, 
+        # 'stack_info': None, 
+        # 'lineno': 97, 
+        # 'funcName': '_log',
+        #  'created': 1725709403.1972203,
+        #  'msecs': 197.0, 
+        # 'relativeCreated': 37105.026721954346,
+        #  'thread': 133472930760384, 
+        # 'threadName': 'Thread-15 (process_request_thread)',
+        #  'processName': 'MainProcess',
+        #  'process': 26, 
+        # 'taskName': None}
+
+
+        entry = LogEntry(
+            message = record.msg,
+            level = record.levelname, #record.levelno,
+            creation_time = datetime.now(),
+
+            name = record.name,
+            pathname = record.pathname,
+            filename = record.filename,
+            module = record.module,
+            lineno = record.lineno,
+            funcName = record.funcName,
+            threadName = record.threadName,
+            processName = record.processName
+        )
+        entry.save()
+
+
+def get_log_level(default=logging.WARN):
+    LOG_LEVEL = os.getenv("LOG_LEVEL")
+    if LOG_LEVEL:
+        return eval("logging." + LOG_LEVEL)
+    return default
+
+
+
+

backend/lib/models.py

@@ -1,7 +1,6 @@
 import os
 from elasticsearch_dsl import Document, InnerDoc, Nested, Date, Integer, Keyword, Float, Long, Text, connections, Object, Boolean
 
-
 class User(Document):
     creation_date = Date()
     email = Keyword()
@@ -29,7 +28,6 @@ class User(Document):
         return super(User, self).save(**kwargs)
 
 
-
 class Chatbot(Document):
     creation_date = Date()
     changed_date = Date()
@@ -63,11 +61,6 @@ class Chatbot(Document):
         return super(Chatbot, self).save(**kwargs)
 
 
-
-
-
-
-
 class Text(Document):
     creation_date = Date()
     creator_id = Keyword()
@@ -84,62 +77,101 @@ class Text(Document):
         return super(Text, self).save(**kwargs)
 
 
-
+class Question(Document):
-#======= Query Log ===========
+    question = Text(index=False, required=True)
-
+    md5 = Keyword()
-
-class Sources(InnerDoc):
-    score = Float()
-    #sourceFileId = Text()
-    sourceType = Text()
-    tags = Text()
-
-    #new fields
-    sourceFileId = Keyword()
-    filename = Keyword()
-    url = Keyword()
-    txt_id = Keyword()
-    page = Integer()
-
-
-
-class QueryLog(Document):
-    answer = Text()
-    question = Text()
-
-    chatbotid = Keyword()
-    durasecs = Float()
-    #inCt = Float()
-    inToks = Long()
-    llm = Text()
-    #outCt = Float()
-    outToks = Long()
-
-    #queryid = Keyword()
-    #rating = Long()
-    #reason = Text()
-    #reasontags = Text()
-    session = Keyword()
-
-    sources = Object(Sources)
-    temperature = Float()
-    #totalCt = Float()
-
-    timest = Date() #timestamp
-    date = Date() #iso date
 
     class Index:
-        name = 'query_log'
+        name = 'question'
         settings = {
             "number_of_shards": 1,
         }
 
     def save(self, ** kwargs):
-        return super(QueryLog, self).save(**kwargs)
+        return super(Question, self).save(**kwargs)
 
 
+class Answer(Document):
+    question_id = Keyword()
+    answer = Text(index=False, required=True)
+    md5 = Keyword()
+
+    class Index:
+        name = 'answer'
+        settings = {
+            "number_of_shards": 1,
+        }
+
+    def save(self, ** kwargs):
+        return super(Answer, self).save(**kwargs)
+
+
+class LogEntry(Document):
+    message = Text(index=False, required=True)
+    level = Keyword() #Integer(required=True)
+    creation_time = Date()
+
+
+    name = Keyword()
+
+    # 'args': ('GET /socket.io/?EIO=4&transport=websocket&sid=MtyTmZQs5IA6DnvhAAAA HTTP/1.1', '200', '-'), 
+
+    pathname = Keyword()
+    #  'pathname': '/usr/local/lib/python3.12/dist-packages/werkzeug/_internal.py', 
+
+    filename = Keyword()
+    # 'filename': '_internal.py', 
+
+    module = Keyword()
+    # 'module': '_internal',
+
+    lineno = Integer(required=True)
+    # 'lineno': 97, 
+
+    funcName = Keyword()
+    # 'funcName': '_log',
+
+
+    #  'created': 1725709403.1972203,
+    #  'msecs': 197.0, 
+
+    threadName = Keyword()
+    # 'threadName': 'Thread-15 (process_request_thread)',
+
+    processName = Keyword()
+    #  'processName': 'MainProcess',
+
+
+
+
+
+    class Index:
+        name = 'logentry'
+        settings = {
+            "number_of_shards": 1,
+        }
+
+    def save(self, ** kwargs):
+        return super(LogEntry, self).save(**kwargs)
+
+
+#======= Query Log ===========
+
+
+#class Sources(InnerDoc):
+#score = Float()
+#tags = Text()
+#filename = Keyword()
+#page = Integer()
+
+
+#----------------------------------------------
 
 def init_indicies():
-    # create the mappings in elasticsearch
+    """
-    for Index in [QueryLog, Chatbot, User, Text]:
+    Create the mappings in elasticsearch
+    """
+    for Index in [LogEntry, Question, Answer, Chatbot, User, Text]:
         Index.init()
+
+

deployment/docker-compose.yml

@@ -32,6 +32,31 @@ services:
       - MINIO_DEFAULT_BUCKETS=defaultbucket
     command: server --console-address ":29001" /data
 
+  neo4j:
+    container_name: ${APP_PREFIX}_neo4j
+    image: neo4j
+    #image: neo4j:3.5
+    #image: neo4j:4.1
+    restart: unless-stopped
+    ports:
+      - 7474:7474
+      - 7687:7687
+    volumes:
+      - ./conf:/conf
+      - ./data:/data
+      - ./import:/import
+      - ./logs:/logs
+      - ./plugins:/plugins
+    environment: 
+      - NEO4J_AUTH=neo4j/your_password
+
+      # Raise memory limits
+      - NEO4J_server_memory_pagecache_size=512M
+      - NEO4J_server_memory_heap_max__size=512M
+
+      - dbms.usage_report.enabled=false
+
+
   elasticsearch:
     container_name: ${APP_PREFIX}_elasticsearch
     image: docker.elastic.co/elasticsearch/elasticsearch:8.11.0
@@ -64,6 +89,10 @@ services:
     restart: always
     ports:
       - "11434:11434"
+    environment:
+      - OLLAMA_NUM_PARALLEL=4
+      - OLLAMA_MAX_LOADED_MODELS=4
+
     volumes:
       - ..:/code
       - ../ollama/ollama:/root/.ollama
@@ -75,6 +104,12 @@ services:
       - /dev/dri
 
 
+
+
+
+
+
+
   #ollama-webui:
   #  container_name: ${APP_PREFIX}_ollama-webui
   #  image: ghcr.io/ollama-webui/ollama-webui:main

deployment/sample.env

@@ -5,6 +5,7 @@ APP_PREFIX=creative_bots
 DEFAULT_USERS=[["user@mail.net", "12345", "user"], ["admin@mail.net", "12345", "admin"]]
 
 #JWT encryption secret:
-SECRET=1234
+SECRET=23A344F670E
-
 
+#WARN INFO FATAL
+LOG_LEVEL=WARN