Web server now has separate Generate and Describe routes

Author: mrseanryan

gpt-workflow-csharp/gpt-workflow-csharp-cli/Builder/DotBuilder.cs

@@ -33,7 +33,7 @@ public string Build()
         sb.AppendLine("digraph G {");
 
         foreach(var node in nodes)
-            sb.AppendLine($"  {GetNodeAsDot(node)}[shape={GetShape(node)}, label=\"{node.Identifier}\" ]");
+            sb.AppendLine($"  {GetNodeAsDot(node)}[shape={GetShape(node)}, label=\"{node.Identifier}\" ];");
 
         foreach(var edge in edges)
         {

main_web_service.py

@@ -1,15 +1,13 @@
 from http.server import BaseHTTPRequestHandler, HTTPServer
-from random import randint
 from urllib.parse import urlparse, parse_qs
+import urllib
 import threading, socket
 
-from prompts_dot_graph_creator import EXPERT_COMMANDS
+from prompts_dot_graph_creator import EXPERT_COMMANDS, GetPromptToDescribeWorkflow, getExpertCommandToCreateDot, getExpertCommandToDescribeDot
 import core
 import config
 import config_web
 
-chain = core.create_command_messages(EXPERT_COMMANDS)
-
 # Python web server with cookie-based session
 # based on https://davidgorski.ca/posts/sessions/
 
@@ -21,7 +19,8 @@
 class MyServer(BaseHTTPRequestHandler):
     def do_GET(self):
         routes = {
-            "/":        self.bot
+            "/generate-dot":        self.bot_generate_dot,
+            "/describe-dot":        self.bot_describe_dot,
         }
         try:
             response = 200
@@ -40,13 +39,23 @@ def do_GET(self):
         self.write(content)
         return
 
-    def bot(self):
+    def bot_describe_dot(self):
+        global prompt_id
         user_prompt = self.parse_query_param("p")
-        command_messages = core.create_command_messages(EXPERT_COMMANDS)
-        rsp = core.execute_prompt(user_prompt, previous_messages, command_messages, prompt_id)
+        user_prompt_wrapped = GetPromptToDescribeWorkflow(user_prompt)
+        command_messages = core.create_command_messages([getExpertCommandToDescribeDot()])
+        rsp = core.execute_prompt(user_prompt_wrapped, previous_messages, command_messages, prompt_id)
         prompt_id += 1
         return rsp
 
+    def bot_generate_dot(self):
+        global prompt_id
+        user_prompt = self.parse_query_param("p")
+        command_messages = core.create_command_messages([getExpertCommandToCreateDot()])
+        rsp = core.execute_prompt(user_prompt, previous_messages, command_messages, prompt_id)
+        prompt_id += 1
+        return rsp["human_output"] + "\n\n======\n\n" + rsp["dot"]
+
     def parse_path(self):
         return urlparse(self.path).path
 
@@ -76,6 +85,9 @@ def start_single_threaded():
     webServer.server_close()
     print("Server stopped.")
 
+def quotify(text):
+    return urllib.parse.quote(text)
+
 def start_multi_threaded():
     # Multi-threaded server, else performance is terrible
     # ref https://stackoverflow.com/questions/46210672/python-2-7-streaming-http-server-supporting-multiple-connections-on-one-port
@@ -108,9 +120,13 @@ def run(self):
                 pass
             httpd.server_close()
 
+    escaped_create_dot = quotify("Create a flow that makes a series of decisions about whether to approve a mortgage application")
+    escaped_dot = quotify('digraph G { Start_Start_1[shape=ellipse, label="Start_1" ];  Decision_Experience[shape=Mdiamond, label="Experience" ];  Decision_Education[shape=Mdiamond, label="Education" ];  Decision_Skills[shape=Mdiamond, label="Skills" ];  End_Recommend[shape=rectangle, label="Recommend" ];  End_Not_Recommend[shape=rectangle, label="Not_Recommend" ];  Start_Start_1 -> Decision_Experience;  Decision_Experience -> Decision_Education [label="true"];  Decision_Experience -> End_Not_Recommend [label="false"];  Decision_Education -> Decision_Skills [label="true"];  Decision_Education -> End_Not_Recommend [label="false"];  Decision_Skills -> End_Recommend [label="true"];  Decision_Skills -> End_Not_Recommend [label="false"];}")}')
+
     print(f"Server started at http://{config_web.HOSTNAME}:{config_web.PORT} - {config_web.WEB_SERVER_THREADS} threads")
     print("Please set the 'p' query parameter to be the user's prompt.")
-    print("example: http://localhost:8083/?p=I%20need%20a%20make%20a%20Car%20Parts%20application")
+    print(f"- generate DOT example: http://localhost:8083/generate-dot?p={escaped_create_dot}")
+    print(f"- describe DOT example: http://localhost:8083/describe-dot?p={escaped_dot}");
     print("[press any key to stop]")
     [Thread(i) for i in range(config_web.WEB_SERVER_THREADS)]
     input("Press ENTER to kill server")

prompts_dot_graph_creator.py

@@ -45,12 +45,18 @@
 - use clear natural language, in a concise, friendly tone.
 """
 
+def getExpertCommandToCreateDot():
+    return Command('create_dot_workflow', create_dot_flowchart__expert_template, "Good for answering questions about creating a workflow in DOT notation")
+
+def getExpertCommandToDescribeDot():
+    return Command('describe_dot_workflow', describe_dot_flowchart__expert_template, "Good for describing a workflow given in DOT notation, summarizing its activity and its general purpose")
+
 # Each expert is a prompt that knows how to handle one type of user input
 EXPERT_COMMANDS = [
-    Command('describe_dot_workflow', describe_dot_flowchart__expert_template, "Good for describing a workflow given in DOT notation, summarizing its activity and its general purpose"),
+    getExpertCommandToDescribeDot(),
     # Placing this last, so that its IMPORTANT message about whitelist is not ignored (LLMs tend to ignore content in middle)
     # An approach like LangChain's MULTI_PROMPT_ROUTER_TEMPLATE would avoid this problem.
-    Command('create_dot_workflow', create_dot_flowchart__expert_template, "Good for answering questions about creating a workflow in DOT notation"),
+    getExpertCommandToCreateDot(),
 ]
 
 def GetPromptToDescribeWorkflow(dotText):

service_dot_parser.py

@@ -38,4 +38,7 @@ def parse_dot_and_return_human(rsp, prompt_id):
     write_dot_to_file(dot_string, prompt_id)
     generate_png_from_dot(dot_string, prompt_id)
 
-    return human_output
+    return {
+        "human_output": human_output,
+        "dot": dot_string
+    }