Initial working version - ouptuts DOT format in response

Author: mrseanryan

README.md

@@ -1,2 +1,260 @@
 # gpt-workflow
-Generate workflows (for flowcharts or low code) via LLM
+Generate workflows (for flowcharts or low code) via LLM.
+
+## Approach: generate DOT notation as a simple format to represent a workflow
+
+The DOT graph format (as used by tools like graphviz) is a simple way to represent a flow chart.
+The DOT script generated by the LLM can be further processed, for example by generating a flow chart image OR by populating some kind of workflow system.
+
+# Example Execution
+
+```
+[[[TEST Simple workflow to decide the animal type]]]
+---
+>> Create a flow that makes a series of decisions about whether to approve a mortgage application
+digraph G {
+
+  // start
+  start [shape=ellipse, label="Start"];
+
+  // decision_credit_score
+  start -> decision_credit_score;
+  decision_credit_score [shape=Mdiamond, label="Credit Score > 700?"];
+
+  // decision_income
+  decision_credit_score -> decision_income;
+  decision_income [shape=Mdiamond, label="Income > $50,000?"];
+
+  // decision_employment
+  decision_income -> decision_employment;
+  decision_employment [shape=Mdiamond, label="Employment > 2 years?"];
+
+  // decision_down_payment
+  decision_employment -> decision_down_payment;
+  decision_down_payment [shape=Mdiamond, label="Down Payment > 20%?"];
+
+  // approve
+  decision_down_payment -> approve;
+  approve [shape=box, label="Approve"];
+
+  // reject
+  decision_credit_score -> reject;
+  reject [shape=box, label="Reject"];
+
+  decision_income -> reject;
+  decision_employment -> reject;
+  decision_down_payment -> reject;
+}
+---
+>> Create a flow that makes a series of decisions about whether to recommend a job interview candidate.
+digraph G {
+
+  // start
+  start [shape=ellipse, label="Start"];
+
+  // decision_experience
+  start -> decision_experience;
+  decision_experience [shape=Mdiamond, label="Has relevant experience?"];
+
+  // decision_education
+  decision_experience -> decision_education;
+  decision_education [shape=Mdiamond, label="Has required education?"];
+
+  // decision_skills
+  decision_education -> decision_skills;
+  decision_skills [shape=Mdiamond, label="Has necessary skills?"];
+
+  // decision_references
+  decision_skills -> decision_references;
+  decision_references [shape=Mdiamond, label="Has positive references?"];
+
+  // recommend
+  decision_references -> recommend;
+  recommend [shape=box, label="Recommend for interview"];
+
+  // reject
+  decision_experience -> reject;
+  reject [shape=box, label="Reject"];
+
+  decision_education -> reject;
+  decision_skills -> reject;
+  decision_references -> reject;
+}
+---
+>> Create a flow that makes a series of decisions about an animal, to decide what kind of animal is it
+digraph G {
+
+  // start
+  start [shape=ellipse, label="Start"];
+
+  // decision_has_feathers
+  start -> decision_has_feathers;
+  decision_has_feathers [shape=Mdiamond, label="Has feathers?"];
+
+  // decision_can_fly
+  decision_has_feathers -> decision_can_fly;
+  decision_can_fly [shape=Mdiamond, label="Can fly?"];
+
+  // decision_has_fins
+  decision_has_feathers -> decision_has_fins;
+  decision_has_fins [shape=Mdiamond, label="Has fins?"];
+
+  // Hawk
+  decision_can_fly -> Hawk;
+  Hawk [shape=box, label="Hawk"];
+
+  // Penguin
+  decision_can_fly -> Penguin;
+  Penguin [shape=box, label="Penguin"];
+
+  // Dolphin
+  decision_has_fins -> Dolphin;
+  Dolphin [shape=box, label="Dolphin"];
+
+  // Bear
+  decision_has_fins -> Bear;
+  Bear [shape=box, label="Bear"];
+}
+[[[TEST Simple workflow adding an item to a list]]]
+---
+>> Create a flow that takes a list and adds an item of the same type
+digraph G {
+
+  // start
+  start [shape=ellipse, label="Start"];
+
+  // add_item
+  add_item [shape=box, label="Add Item"];
+
+  // end
+  end [shape=ellipse, label="End"];
+
+  // start -> add_item
+  start -> add_item;
+
+  // add_item -> end
+  add_item -> end;
+
+  add_item [shape=box, label="Add Item"];
+}
+---
+>> Create a flow that takes two lists and concatenates them
+digraph G {
+
+  // start
+  start [shape=ellipse, label="Start"];
+
+  // concatenate_lists
+  concatenate_lists [shape=box, label="Concatenate Lists"];
+
+  // end
+  end [shape=ellipse, label="End"];
+
+  // start -> concatenate_lists
+  start -> concatenate_lists;
+
+  // concatenate_lists -> end
+  concatenate_lists -> end;
+
+  concatenate_lists [shape=box, label="Concatenate Lists"];
+}
+---
+>> Create a flow that takes a list and an object. Call another flow to get a boolean result. If the boolean is true, then add the item to the list.
+digraph G {
+
+  // start
+  start [shape=ellipse, label="Start"];
+
+  // call_flow
+  call_flow [shape=box, label="Call Flow"];
+
+  // decision_boolean
+  decision_boolean [shape=diamond, label="Boolean Result?"];
+
+  // add_item
+  add_item [shape=box, label="Add Item"];
+
+  // end
+  end [shape=ellipse, label="End"];
+
+  // start -> call_flow
+  start -> call_flow;
+
+  // call_flow -> decision_boolean
+  call_flow -> decision_boolean;
+
+  // decision_boolean -> add_item [label="true"];
+  decision_boolean -> add_item [label="true"];
+
+  // decision_boolean -> end [label="false"];
+  decision_boolean -> end [label="false"];
+
+  // add_item -> end
+  add_item -> end;
+
+  call_flow [shape=box, label="Call Flow"];
+}
+[[[TEST Irrelevant prompts]]]
+---
+>> what is 2 + 5 divided by 10 ?
+I'm sorry, but I can only assist with questions related to creating a flow chart.
+---
+>> Who won the battle of Agincourt, and why was it fought?
+I'm sorry, but I can only assist with questions related to creating a flow chart.
+---
+>> What is my favourite color?
+I'm sorry, but I don't have access to personal information.
+```
+
+## Dependencies
+
+- Requires an LLM - by default, uses OpenAI's ChatGPT.
+- Python 3
+- [graphviz](https://www.graphviz.org/#download)
+
+## Usage
+
+To use as a CLI (Command Line Interface) REPL (Read-Eval-Print Loop) prompt:
+```go.sh```
+
+or to use as a web server:
+
+```go_web.sh```
+
+For the web server, you need to pass the user prompt as GET query parameter 'p'.
+
+Example:
+
+- http://localhost:8083/?p=I%20need%20a%20make%20a%20Car%20Parts%20application
+
+So, another application can use the web server to send in natural language prompts from the user, and receive response in the graphviz DOT format.
+
+The other application can then generate an image or some kind of workflow, from the DOT script.
+
+## Set up
+
+```
+pip3 install --upgrade openai pydot
+```
+
+Set environment variable with your OpenAI key:
+
+```
+export OPENAI_API_KEY="xxx"
+```
+
+Add that to your shell initializing script (`~/.zprofile` or similar)
+
+Load in current terminal:
+
+```
+source ~/.zprofile
+```
+
+## Test
+
+`test.sh`
+
+or
+
+`python test.py`

command.py

@@ -0,0 +1,5 @@
+class Command:
+    def __init__(self, name, expert_template, description) -> None:
+        self.name = name
+        self.expert_template = expert_template
+        self.description = description

config.py

@@ -0,0 +1 @@
+is_debug = False

config__database_schema_creator.py

@@ -0,0 +1,42 @@
+from command import Command
+
+# note: Instead of this steps-in-single-prompt approach, we could have an inter-dependent chain, to collect info about the app, THEN try to generate.
+# BUT the step-by-step approach works really well, at least with Chat-GPT3.5 Turbo.
+
+create_dot_flowchart__expert_template = """You are Workflow Creator Bot, a bot that knows how to create a simple DOT format flow chart.
+You are great at answering questions about creating and altering a flow chart.
+
+When you don't know the answer to a question, do not answer.
+
+You are an AI assistant to assist an application developer with the creation of the flow chart via natural language input.
+
+The output MUST be in DOT format as used by the graphviz tool only, based on the following example:
+```
+digraph G {
+
+  // decision_has_feathers
+  start -> decision_has_feathers;
+  decision_has_feathers -> decision_can_fly
+  decision_has_feathers -> decision_has_fins
+  
+  // decision_can_fly
+  decision_can_fly -> Hawk
+  decision_can_fly -> Penguin
+
+  // decision_has_fins
+  decision_has_fins -> Dolphin
+  decision_has_fins -> Bear
+
+  decision_has_feathers [shape=Mdiamond, label="Has feathers?"];
+  decision_can_fly [shape=Mdiamond, label="Can fly?"];
+  decision_has_fins [shape=Mdiamond, label="Has fins?"];
+}
+```
+
+IMPORTANT: Only output valid DOT format as used by the graphviz tool.
+"""
+
+# Each expert is a prompt that knows how to handle one type of user input
+EXPERT_COMMANDS = [
+    Command('create_dot_workflow', create_dot_flowchart__expert_template, "Good for answering questions about creating a workflow in DOT notation")
+]

config_web.py

@@ -0,0 +1,5 @@
+HOSTNAME = "localhost"
+PORT = 8083
+WEB_SERVER_THREADS = 100
+
+is_debug = False

core.py

@@ -0,0 +1,26 @@
+import service_chat
+
+# DEV note: NOT using langchain.
+# Tried using langchain but it's validation gets in the way of more complex prompts.
+# And seems simpler to code direct, not via a complicated framework.
+
+def create_command_messages(expert_commands):
+    messages = []
+    for command in expert_commands:
+        messages.append({'role':'system', 'content': command.expert_template })
+
+    return messages
+
+def execute_prompt(user_prompt, previous_messages, command_messages):
+    # TODO: Route to the right 'expert' chain
+    # Falls back to the default chain, which means sending the plain user prompt to the LLM
+
+    user_message = {'role':'user', 'content': user_prompt }
+
+    messages = command_messages + previous_messages + [user_message]
+
+    rsp = service_chat.send_prompt_messages(messages)
+
+    previous_messages.append(user_message)
+    previous_messages.append({'role':'assistant', 'content': rsp })
+    return rsp

go.sh

@@ -0,0 +1 @@
+python -W "ignore" main_cli.py

go_web.sh

@@ -0,0 +1 @@
+python3 -W "ignore" main_web_service.py