add initial prompt stuff

Makefile

@@ -0,0 +1,14 @@
+.PHONY: format lint tests
+
+format:
+	black .
+	isort .
+
+lint: 
+	black . --check
+	isort . --check
+	flake8 .
+	mypy .
+
+tests:
+	pytest tests

langchain/formatting.py

@@ -0,0 +1,21 @@
+from string import Formatter
+
+
+class StrictFormatter(Formatter):
+    """A subclass of formatter that checks for extra keys."""
+
+    def check_unused_args(self, used_args, args, kwargs):
+        extra = set(kwargs).difference(used_args)
+        if extra:
+            raise KeyError(extra)
+
+    def vformat(self, format_string, args, kwargs):
+        if len(args) > 0:
+            raise ValueError(
+                "No arguments should be provided, "
+                "everything should be passed as keyword arguments."
+            )
+        return super().vformat(format_string, args, kwargs)
+
+
+formatter = StrictFormatter()

langchain/prompt.py

@@ -0,0 +1,27 @@
+"""Base schema types."""
+from typing import Dict, List
+
+from pydantic import BaseModel, Extra, root_validator
+
+from langchain.formatting import formatter
+
+
+class Prompt(BaseModel):
+    """Schema to represent a prompt for an LLM."""
+
+    input_variables: List[str]
+    template: str
+
+    class Config:
+        extra = Extra.forbid
+
+    @root_validator()
+    def template_is_valid(cls, values: Dict) -> Dict:
+        input_variables = values["input_variables"]
+        template = values["template"]
+        dummy_inputs = {input_variable: "foo" for input_variable in input_variables}
+        try:
+            formatter.format(template, **dummy_inputs)
+        except KeyError:
+            raise ValueError("Invalid prompt schema.")
+        return values

tests/unit_tests/data/prompts/prompt_extra_args.json

@@ -0,0 +1,5 @@
+{
+  "input_variables": ["foo"],
+  "template": "This is a {foo} test.",
+  "bad_var": 1
+}

tests/unit_tests/data/prompts/prompt_missing_args.json

@@ -0,0 +1,3 @@
+{
+  "input_variables": ["foo"]
+}

tests/unit_tests/data/prompts/simple_prompt.json

@@ -0,0 +1,4 @@
+{
+  "input_variables": ["foo"],
+  "template": "This is a {foo} test."
+}

tests/unit_tests/test_formatting.py

@@ -0,0 +1,22 @@
+import pytest
+
+from langchain.formatting import formatter
+
+
+def test_valid_formatting():
+    template = "This is a {foo} test."
+    output = formatter.format(template, foo="good")
+    expected_output = "This is a good test."
+    assert output == expected_output
+
+
+def test_does_not_allow_args():
+    template = "This is a {} test."
+    with pytest.raises(ValueError):
+        formatter.format(template, "good")
+
+
+def test_does_not_allow_extra_kwargs():
+    template = "This is a {foo} test."
+    with pytest.raises(KeyError):
+        formatter.format(template, foo="good", bar="oops")

tests/unit_tests/test_schema.py

@@ -0,0 +1,32 @@
+import pytest
+
+from langchain.prompt import Prompt
+
+
+def test_prompt_valid():
+    template = "This is a {foo} test."
+    input_variables = ["foo"]
+    prompt = Prompt(input_variables=input_variables, template=template)
+    assert prompt.template == template
+    assert prompt.input_variables == input_variables
+
+
+def test_prompt_missing_input_variables():
+    template = "This is a {foo} test."
+    input_variables = []
+    with pytest.raises(ValueError):
+        Prompt(input_variables=input_variables, template=template)
+
+
+def test_prompt_extra_input_variables():
+    template = "This is a {foo} test."
+    input_variables = ["foo", "bar"]
+    with pytest.raises(ValueError):
+        Prompt(input_variables=input_variables, template=template)
+
+
+def test_prompt_wrong_input_variables():
+    template = "This is a {foo} test."
+    input_variables = ["bar"]
+    with pytest.raises(ValueError):
+        Prompt(input_variables=input_variables, template=template)