added readme and setup for install

ProjectCli/modules/review.py

@@ -0,0 +1,77 @@
+import os
+
+from ollama import ChatResponse, chat
+
+
+def RunCommand(command) -> str:
+    """Runs a command on the HOST os"""
+    """
+  Args:
+      command; Parsed command the LLM sends
+
+  Returns:
+      Output of command
+  """
+    return os.popen(command).read()
+
+
+class review:
+    def __init__(self) -> None:
+        pass
+
+    def generateReview(self, providedModel, platform, markdown) -> str:
+        # save the messageHistory while the model is reviewing the project
+        messageHistory = [
+            {
+                "role": "system",
+                "content": f"""You are going to review a project the user made based on a MD file.
+- The user is currently on the platform; {platform}
+- You will be allowed using tool calling to execute commands directly on the OS machine.
+- You are NOT allowed to search up the web, do curl requests, modify files, or harm the device. The tool calling is intended purely for read only access
+- Upon checking and validating the code generate your output (message content) in markdown formatted text as this will be directly injected into A markdown file
+- do NOT use markdown blocks or anything else that would break the MD file in your output.
+- review the user based on the score provided in the instuctions the user is going to provide. If not provided make up your own based on the instuctions
+- When creating the markdown file always include the following things; Introduction, Review, Evaluation, Score, Rating, Comments, Extra information (if needed)
+- when your evalution the project use the terminal (tool calling) to inspect files and read them.
+- be VERY critical when reviewing the project. The user has to learn fropm this experience.
+
+
+-- MARKDOWN FILE --
+{markdown}
+                """,
+            },
+            {
+                "role": "user",
+                "content": "Using tool calling and instructions provided in sys prompt, Complete the review.",
+            },
+        ]
+
+        while True:
+            # print("generate")
+            response: ChatResponse = chat(
+                model=providedModel,
+                messages=messageHistory,
+                tools=[RunCommand],
+            )
+            # print("Content: ", response.message.content)
+            if response.message.tool_calls:
+                print("attempted tool calling")
+                for tool in response.message.tool_calls:
+                    print("-- TOOL CALLING RAN --")
+                    print(
+                        "- The LLM is currently running tests on your computer in order to generate a review -"
+                    )
+
+                    print(f"Command; {tool.function.arguments}")
+                    result = RunCommand(**tool.function.arguments)
+                    print(f"Result: {result}")
+                    messageHistory.append(
+                        {
+                            "role": "tool",
+                            "tool_name": tool.function.name,
+                            "content": str(result),
+                        }
+                    )
+                    print("-- END OF TOOL CALL --")
+            else:
+                return response["message"]["content"]