Separate prompt builder concept from adapter.

.vscode/settings.json

@@ -4,7 +4,8 @@
     "editor.defaultFormatter": "charliermarsh.ruff",
     "editor.codeActionsOnSave": {
       "source.organizeImports": "explicit"
-    }
+    },
+    "editor.tabSize": 4,
   },
   "eslint.validate": ["javascript", "javascriptreact", "svelte"],
   // Svelte, JS, TS files
@@ -24,5 +25,5 @@
     "**/.ruff_cache": true,
     "**/dist": true,
     "**/node_modules": true
-  }
+  },
 }

libs/core/kiln_ai/adapters/base_adapter.py

@@ -1,7 +1,18 @@
 from abc import ABCMeta, abstractmethod
 
+from kiln_ai.datamodel.models import Task
+
 
 class BaseAdapter(metaclass=ABCMeta):
     @abstractmethod
     async def run(self, input: str) -> str:
         pass
+
+
+class BasePromptBuilder(metaclass=ABCMeta):
+    def __init__(self, task: Task):
+        self.task = task
+
+    @abstractmethod
+    def build_prompt(self, input: str) -> str:
+        pass

libs/core/kiln_ai/adapters/prompt_adapters.py → libs/core/kiln_ai/adapters/langchain_adapters.py

@@ -1,21 +1,22 @@
 import json
-from abc import ABCMeta, abstractmethod
 
 import kiln_ai.datamodel.models as models
+from kiln_ai.adapters.prompt_builders import SimplePromptBuilder
 from langchain_core.language_models.chat_models import BaseChatModel
 from langchain_core.messages.base import BaseMessage
 
-from .base_adapter import BaseAdapter
+from .base_adapter import BaseAdapter, BasePromptBuilder
 from .ml_model_list import langchain_model_from
 
 
-class BaseLangChainPromptAdapter(BaseAdapter, metaclass=ABCMeta):
+class LangChainPromptAdapter(BaseAdapter):
     def __init__(
         self,
         kiln_task: models.Task,
         custom_model: BaseChatModel | None = None,
         model_name: str | None = None,
         provider: str | None = None,
+        prompt_builder: BasePromptBuilder | None = None,
     ):
         self.kiln_task = kiln_task
         self.__is_structured = False
@@ -46,16 +47,15 @@ def __init__(
                 output_schema, include_raw=True
             )
             self.__is_structured = True
-
-    @abstractmethod
-    def build_prompt(self) -> str:
-        pass
+        if prompt_builder is None:
+            self.prompt_builder = SimplePromptBuilder(kiln_task)
+        else:
+            self.prompt_builder = prompt_builder
 
     # TODO: don't just append input to prompt
     async def run(self, input: str) -> str:
         # TODO cleanup
-        prompt = self.build_prompt()
-        prompt += f"\n\n{input}"
+        prompt = self.prompt_builder.build_prompt(input)
         response = self.model.invoke(prompt)
         if self.__is_structured:
             if (
@@ -74,17 +74,3 @@ async def run(self, input: str) -> str:
             if not isinstance(text_content, str):
                 raise RuntimeError(f"response is not a string: {text_content}")
             return text_content
-
-
-class SimplePromptAdapter(BaseLangChainPromptAdapter):
-    def build_prompt(self) -> str:
-        base_prompt = self.kiln_task.instruction
-
-        # TODO: this is just a quick version. Formatting and best practices TBD
-        if len(self.kiln_task.requirements()) > 0:
-            base_prompt += "\n\nYou should requect the following requirements:\n"
-            # iterate requirements, formatting them in numbereed list like 1) task.instruction\n2)...
-            for i, requirement in enumerate(self.kiln_task.requirements()):
-                base_prompt += f"{i+1}) {requirement.instruction}\n"
-
-        return base_prompt

libs/core/kiln_ai/adapters/prompt_builders.py

@@ -0,0 +1,19 @@
+from kiln_ai.adapters.base_adapter import BasePromptBuilder
+
+
+class SimplePromptBuilder(BasePromptBuilder):
+    def build_prompt(self, input: str) -> str:
+        base_prompt = self.task.instruction
+
+        # TODO: this is just a quick version. Formatting and best practices TBD
+        if len(self.task.requirements()) > 0:
+            base_prompt += (
+                "\n\nYour response should respectthe following requirements:\n"
+            )
+            # iterate requirements, formatting them in numbereed list like 1) task.instruction\n2)...
+            for i, requirement in enumerate(self.task.requirements()):
+                base_prompt += f"{i+1}) {requirement.instruction}\n"
+
+        # TODO: should be another message
+        base_prompt += f"\n\nThe input is:\n{input}"
+        return base_prompt

libs/core/kiln_ai/adapters/test_prompt_adaptors.py

@@ -3,8 +3,8 @@
 
 import kiln_ai.datamodel.models as models
 import pytest
+from kiln_ai.adapters.langchain_adapters import LangChainPromptAdapter
 from kiln_ai.adapters.ml_model_list import built_in_models, ollama_online
-from kiln_ai.adapters.prompt_adapters import SimplePromptAdapter
 from langchain_core.language_models.fake_chat_models import FakeListChatModel
 
 
@@ -62,7 +62,7 @@ async def test_amazon_bedrock(tmp_path):
 async def test_mock(tmp_path):
     task = build_test_task(tmp_path)
     mockChatModel = FakeListChatModel(responses=["mock response"])
-    adapter = SimplePromptAdapter(task, custom_model=mockChatModel)
+    adapter = LangChainPromptAdapter(task, custom_model=mockChatModel)
     answer = await adapter.run("You are a mock, send me the response!")
     assert "mock response" in answer
 
@@ -116,7 +116,8 @@ async def run_simple_test(tmp_path: Path, model_name: str, provider: str | None
 
 
 async def run_simple_task(task: models.Task, model_name: str, provider: str):
-    adapter = SimplePromptAdapter(task, model_name=model_name, provider=provider)
+    adapter = LangChainPromptAdapter(task, model_name=model_name, provider=provider)
+
     answer = await adapter.run(
         "You should answer the following question: four plus six times 10"
     )

libs/core/kiln_ai/adapters/test_structured_output.py

@@ -3,11 +3,11 @@
 
 import kiln_ai.datamodel.models as models
 import pytest
+from kiln_ai.adapters.langchain_adapters import LangChainPromptAdapter
 from kiln_ai.adapters.ml_model_list import (
     built_in_models,
     ollama_online,
 )
-from kiln_ai.adapters.prompt_adapters import SimplePromptAdapter
 from kiln_ai.datamodel.test_models import json_joke_schema
 
 
@@ -89,7 +89,7 @@ def build_structured_output_test_task(tmp_path: Path):
 
 async def run_structured_output_test(tmp_path: Path, model_name: str, provider: str):
     task = build_structured_output_test_task(tmp_path)
-    a = SimplePromptAdapter(task, model_name=model_name, provider=provider)
+    a = LangChainPromptAdapter(task, model_name=model_name, provider=provider)
     result = await a.run("Cows")  # a joke about cows
     parsed = json.loads(result)
     assert parsed["setup"] is not None