00-00

neo4jRAG.py

@@ -1,7 +1,260 @@
+import os
+import random
+import requests
+import json
 import gradio as gr
 
-def greet(name):
-    return "Hello " + name + "!!"
+from dotenv import load_dotenv
+from textwrap import dedent
+from snowflake import connector as sf
+from neo4j import GraphDatabase as nj
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+load_dotenv()
+
+sf__user = os.getenv('sf__user')
+sf__passkey = os.getenv('sf__passkey')
+sf__identity = os.getenv('sf__identity')
+sf__warehouse = os.getenv('sf__warehouse')
+sf__database = os.getenv('sf__database')
+sf__schema = os.getenv('sf__schema')
+
+nj__uri = os.getenv('nj__uri')
+nj__username = os.getenv('nj__username')
+nj__password = os.getenv('nj__password')
+
+sf__connection = sf.connect(user=sf__user, password=sf__passkey, account=sf__identity, warehouse=sf__warehouse, database=sf__database, schema=sf__schema)
+sf__cursor = sf__connection.cursor()
+sf__dt_updated__table = sf__cursor.execute(f"SELECT * FROM INDIGO_AIRLINES__CUSTOMER_DATA").fetch_pandas_all().to_dict('records')
+sf__cursor.close()
+sf__connection.close()
+
+nj__driver = nj.driver(nj__uri, auth=(nj__username, nj__password))
+nj__session = nj__driver.session()
+
+nj__n_nodes = nj__session.run("""MATCH (n) RETURN count(n)""", database_ = 'neo4j').single()[0]
+
+
+if(): pass
+
+elif(nj__n_nodes == 0):
+
+  prmpt_sys__sql2cql = \
+  f"""
+  name :: neo4jRAG : sql2cql
+
+  task :: output the 'cypher query' consisting of {{nodes, relationships, property keys}} as per the input data.
+
+  constraints ::
+  - strictly ensure the syntax of the cypher query based on the 'neo4j' syntax.
+  - strictly ensure to hide intro, outro and code blocks.
+
+  structure ::
+    input :: [{{...}},...,{{...}}]
+    output :: <cypher query(s) seperated with ';'>
+  """
+
+  url = "https://bedrock-runtime.us-east-1.amazonaws.com/model/us.amazon.nova-premier-v1:0/converse"
+
+  lst_itr__sql2cql = []
+  for itr in range(0, len(sf__dt_updated__table), 10):
+    payload = {
+        "messages": [{"role": "user","content": [{"text": str(sf__dt_updated__table[itr:itr+10])}]}],
+        "system" : [{"text" : prmpt_sys__sql2cql}]
+    }
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {os.getenv('aws__bedrock')}"
+    }
+    response = requests.request("POST", url, json=payload, headers=headers)
+
+    lst_itr__sql2cql.append(json.loads(response.text)['output']['message']['content'][0]['text'].replace('\n', ';'))
+
+  str_qry__sql2cql = ";".join(lst_itr__sql2cql)
+  lst_qry__sql2cql = list(filter(None, str_qry__sql2cql.split(";")))
+
+  itr_qry__sql = 0
+  itr_qry__cql = 200
+
+  while itr_qry__cql>0 and itr_qry__sql<len(lst_qry__sql2cql):
+    try:
+      nj__session.run(lst_qry__sql2cql[itr_qry__sql])
+      itr_qry__cql -= 1
+      itr_qry__sql += 1
+    except:
+      itr_qry__sql += 1
+
+else: pass
+
+nj__schema = nj__session.run\
+ (
+  """
+  CALL db.labels() YIELD label
+  WITH collect(label) AS lst__labels
+  CALL db.relationshipTypes() YIELD relationshipType
+  WITH lst__labels, collect(relationshipType) AS lst__relationshipTypes
+  CALL db.propertyKeys() YIELD propertyKey
+  WITH lst__labels, lst__relationshipTypes, collect(propertyKey) AS lst__propertyKeys
+  RETURN 
+  {
+    nj__nodes: lst__labels,
+    nj__relationships: lst__relationshipTypes,
+    nj__propertykeys: lst__propertyKeys
+  } 
+  AS nj__schema
+  """
+).data()[-1]
+
+nj__plot = nj__session.run\
+ (
+  """
+MATCH (n)-[r]->(m)
+WITH labels(n) AS sourceLabels,
+     type(r) AS relType,
+     labels(m) AS targetLabels,
+     keys(n) AS nodeProps,
+     keys(r) AS relProps
+WHERE size(nodeProps) > 0 OR size(relProps) > 0
+RETURN DISTINCT
+       sourceLabels[0] AS fromNode,
+       relType,
+       targetLabels[0] AS toNode,
+       nodeProps AS sourceProperties,
+       relProps AS relationshipProperties
+ORDER BY fromNode, relType, toNode  
+"""
+ ).data()
+
+### ===================================================================================================================== ###
+
+# prmpt_sys__llm_txt2txt = \
+#   f"""
+#   name :: neo4jRAG : llm_txt2txt
+
+#   task :: output the 8 cypher queries, prompts based on the 'nj__schema' and 'nj__plot'.
+#   constraints ::
+#   - strictly ensure the syntax of the cypher query based on the 'neo4j' syntax.
+#   - strictly ensure to hide intro, outro and code blocks.
+
+#   structure ::
+#     output :: [{{'prompt' : '...', 'query' : '...'}}]
+
+#   source ::
+#     nj__schema :: {nj__schema}
+#     nj__plot :: {nj__plot}
+#   """
+
+# url = "https://bedrock-runtime.us-east-1.amazonaws.com/model/us.amazon.nova-premier-v1:0/converse"
+
+# payload = {
+#       "messages": [{"role": "user","content": [{"text": " "}]}],
+#       "system" : [{"text" : prmpt_sys__llm_txt2txt}],
+#       "temperature": 0.6,
+#       "topP" : 0.2
+# }
+# headers = {
+#     "Content-Type": "application/json",
+#     "Authorization": f"Bearer {os.getenv('aws__bedrock')}"
+# }
+# response = requests.request("POST", url, json=payload, headers=headers)
+
+# lst_dct__txt2txt = eval(json.loads(response.text)['output']['message']['content'][0]['text'])
+# for itr in lst_dct__txt2txt:
+#   try:
+#     print()
+#     print(itr['prompt'])
+#     print(itr['query'])
+#     print(nj__session.run(itr['query']).data())
+#     print()
+#   except:
+#     pass
+
+### ===================================================================================================================== ###
+
+
+def func__neo4jRAG(args__msg, args__hstry):
+
+    query = args__msg
+  
+    prmpt_sys__txt2cql = \
+    f"""
+    name :: neo4jRAG : txt2sql
+
+    task :: output the cypher query as per the input prompt based on the 'nj__schema' and 'nj__plot'.
+    constraints ::
+    - strictly ensure the syntax of the cypher query based on the 'neo4j' syntax.
+    - strictly ensure to hide intro, outro and code blocks.
+
+    structure ::
+        input :: <prompt text>
+        output :: <cypher query>
+
+    source ::
+        nj__schema :: {nj__schema}
+        nj__plot :: {nj__plot}
+    """
+
+    url = "https://bedrock-runtime.us-east-1.amazonaws.com/model/us.amazon.nova-premier-v1:0/converse"
+
+    payload = {
+        "messages": [{"role": "user","content": [{"text": f"{query}"}]}],
+        "system" : [{"text" : prmpt_sys__txt2cql}],
+        "temperature": 0.6,
+        "topP" : 0.2
+    }
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {os.getenv('aws__bedrock')}"
+    }
+    response = requests.request("POST", url, json=payload, headers=headers)
+
+    sf_nj__sql2cql = json.loads(response.text)['output']['message']['content'][0]['text']
+    # print(nj__session.run(io_nj__txt2sql).data())
+
+
+    prmpt_sys__cql2txt = \
+    f"""
+    name :: neo4jRAG : cql2txt
+
+    task :: output the info as per the input query based on the data.
+    constraints ::
+    - strictly ensure the overall summarized information based on the data in terms of data analysis in the business perspective with accurate numerical information in the structure of a paragraph.
+    - strictly ensure that the numericals are as per indian standards.
+    - strictly ensure to hide intro, outro and code blocks.
+
+    structure ::
+        input :: <prmpt>
+        output :: <info>
+
+    source ::
+        data :: {sf_nj__sql2cql}
+    """
+
+    url = "https://bedrock-runtime.us-east-1.amazonaws.com/model/us.amazon.nova-premier-v1:0/converse"
+
+    payload = {
+        "messages": [{"role": "user","content": [{"text": f"{query}"}]}],
+        "system" : [{"text" : prmpt_sys__cql2txt}],
+        "temperature": 0.6,
+        "topP" : 0.2,
+        "maxTokens" : 8888
+    }
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {os.getenv('aws__bedrock')}"
+    }
+    response = requests.request("POST", url, json=payload, headers=headers)
+    info = json.loads(response.text)['output']['message']['content'][0]['text']
+
+    queryinfo = \
+        f"""
+        {info}
+        <br>
+        ```cypher
+        {sf_nj__sql2cql.replace("\n", ";")}
+        ```
+        """
+    return dedent(queryinfo)
+
+
+var__neo4jRAG = gr.ChatInterface(func__neo4jRAG, title="neo4jRAG")
+var__neo4jRAG.launch()

requirements.txt

@@ -1,6 +1,5 @@
 python-dotenv==1.2.1
 requests==2.32.5
-pandas==2.3.3
 neo4j==6.0.3
-snowflake-connector-python==4.1.0
+snowflake-connector-python[pandas]==4.1.0
 gradio==6.0.0