refactor: organize LLM chat as a reusable package

llm/__init__.py

@@ -0,0 +1,11 @@
+from .client import LLM
+from .models import LLMResponse, LLMUsage
+from .settings import Settings, load_settings
+
+__all__ = [
+    "LLM",
+    "LLMResponse",
+    "LLMUsage",
+    "Settings",
+    "load_settings",
+]

llm/chat.py

@@ -0,0 +1,87 @@
+from openai import OpenAIError
+
+from .client import LLM
+from .console import print_usage, read_input
+from .settings import load_settings
+
+
+def trim_history(messages, keep_count):
+    """Keep the system message and the latest completed exchanges."""
+    if keep_count < 0:
+        raise ValueError("MEMORY_KEEP_COUNT must be 0 or greater")
+
+    if keep_count == 0:
+        return messages[:1]
+
+    conversation = messages[1:]
+    conversation_start = -(keep_count * 2)
+    return messages[:1] + conversation[conversation_start:]
+
+
+def run_chat():
+    try:
+        settings = load_settings()
+    except ValueError as error:
+        print(f"Configuration error: {error}")
+        return 1
+
+    model = LLM(settings)
+
+    while True:
+        system_rule = read_input("System rule: ")
+
+        if system_rule is None:
+            return 0
+
+        if system_rule:
+            break
+
+        print("Please enter a system rule.")
+
+    messages = [
+        {
+            "role": "system",
+            "content": system_rule,
+        }
+    ]
+
+    while True:
+        user_input = read_input("\nYou: ")
+
+        if user_input is None or user_input.lower() in {"exit", "quit"}:
+            return 0
+
+        if not user_input:
+            print("Please enter a message.")
+            continue
+
+        messages.append(
+            {
+                "role": "user",
+                "content": user_input,
+            }
+        )
+
+        try:
+            response = model.complete(
+                messages,
+                on_text=lambda text: print(text, end="", flush=True),
+            )
+        except OpenAIError as error:
+            messages.pop()
+            print(f"\nRequest failed: {error}")
+            continue
+
+        if response.finish_reason == "length":
+            print("Warning: the answer may be incomplete.")
+
+        if settings.show_usage and response.usage:
+            print_usage(response.usage)
+
+        messages.append(
+            {
+                "role": "assistant",
+                "content": response.content,
+            }
+        )
+        messages = trim_history(messages, settings.memory_keep_count)

llm.py → llm/client.py

@@ -1,7 +1,7 @@
-from dataclasses import dataclass
-
 from openai import OpenAI
 
+from .models import LLMResponse, LLMUsage
+
 
 class LLM:
     def __init__(self, settings, client=None):
@@ -91,20 +91,4 @@ def complete(self, messages, on_text=None):
         return self.build_response(answer, finish_reason, usage)
 
 
-@dataclass(frozen=True)
-class LLMUsage:
-    prompt_tokens: int
-    completion_tokens: int
-    total_tokens: int
-    reasoning_tokens: int | None = None
-    cost: float | None = None
-
-
-@dataclass(frozen=True)
-class LLMResponse:
-    content: str
-    finish_reason: str | None
-    usage: LLMUsage | None = None
-
-
 __all__ = ["LLM", "LLMResponse", "LLMUsage"]

llm/models.py

@@ -0,0 +1,20 @@
+from dataclasses import dataclass
+
+
+@dataclass(frozen=True)
+class LLMUsage:
+    prompt_tokens: int
+    completion_tokens: int
+    total_tokens: int
+    reasoning_tokens: int | None = None
+    cost: float | None = None
+
+
+@dataclass(frozen=True)
+class LLMResponse:
+    content: str
+    finish_reason: str | None
+    usage: LLMUsage | None = None
+
+
+__all__ = ["LLMResponse", "LLMUsage"]

main.py

@@ -1,91 +1,4 @@
-from openai import OpenAIError
-
-from llm import LLM
-from settings import load_settings
-from utils import print_usage, read_input
-
-
-def trim_history(messages, keep_count):
-    """Keep the system message and the latest completed exchanges."""
-    if keep_count < 0:
-        raise ValueError("MEMORY_KEEP_COUNT must be 0 or greater")
-
-    if keep_count == 0:
-        return messages[:1]
-
-    conversation = messages[1:]
-    conversation_start = -(keep_count * 2)
-    return messages[:1] + conversation[conversation_start:]
-
-
-def main():
-    try:
-        settings = load_settings()
-    except ValueError as error:
-        print(f"Configuration error: {error}")
-        return 1
-
-    model = LLM(settings)
-
-    while True:
-        system_rule = read_input("System rule: ")
-
-        if system_rule is None:
-            return 0
-
-        if system_rule:
-            break
-
-        print("Please enter a system rule.")
-
-    messages = [
-        {
-            "role": "system",
-            "content": system_rule,
-        }
-    ]
-
-    while True:
-        user_input = read_input("\nYou: ")
-
-        if user_input is None or user_input.lower() in {"exit", "quit"}:
-            return 0
-
-        if not user_input:
-            print("Please enter a message.")
-            continue
-
-        messages.append(
-            {
-                "role": "user",
-                "content": user_input,
-            }
-        )
-
-        try:
-            response = model.complete(
-                messages,
-                on_text=lambda text: print(text, end="", flush=True),
-            )
-        except OpenAIError as error:
-            messages.pop()
-            print(f"\nRequest failed: {error}")
-            continue
-
-        if response.finish_reason == "length":
-            print("Warning: the answer may be incomplete.")
-
-        if settings.show_usage and response.usage:
-            print_usage(response.usage)
-
-        messages.append(
-            {
-                "role": "assistant",
-                "content": response.content,
-            }
-        )
-        messages = trim_history(messages, settings.memory_keep_count)
-
+from llm.chat import run_chat
 
 if __name__ == "__main__":
-    raise SystemExit(main())
+    raise SystemExit(run_chat())

tests/test_main.py

@@ -1,6 +1,6 @@
 import unittest
 
-from main import trim_history
+from llm.chat import trim_history
 
 
 class TrimHistoryTests(unittest.TestCase):

tests/test_settings.py

@@ -1,6 +1,6 @@
 import unittest
 
-from settings import Settings
+from llm.settings import Settings
 
 
 class SettingsTests(unittest.TestCase):