partial variables (langchain-ai#1308)

Author: hwchase17

docs/modules/prompts/examples/partial.ipynb

@@ -0,0 +1,184 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "id": "9355a547",
+ "metadata": {},
+ "source": [
+ "# Partial Prompt Templates\n",
+ "\n",
+ "A prompt template is a class with a `.format` method which takes in a key-value map and returns a string (a prompt) to pass to the language model. Like other methods, it can make sense to \"partial\" a prompt template - eg pass in a subset of the required values, as to create a new prompt template which expects only the remaining subset of values.\n",
+ "\n",
+ "LangChain supports this in two ways: we allow for partially formatted prompts (1) with string values, (2) with functions that return string values. These two different ways support different use cases. In the documentation below we go over the motivations for both use cases as well as how to do it in LangChain.\n",
+ "\n",
+ "## Partial With Strings\n",
+ "\n",
+ "One common use case for wanting to partial a prompt template is if you get some of the variables before others. For example, suppose you have a prompt template that requires two variables, `foo` and `baz`. If you get the `foo` value early on in the chain, but the `baz` value later, it can be annoying to wait until you have both variables in the same place to pass them to the prompt template. Instead, you can partial the prompt template with the `foo` value, and then pass the partialed prompt template along and just use that. Below is an example of doing this:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "643af5da",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from langchain.prompts import PromptTemplate"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "4080d8d7",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "foobaz\n"
+ ]
+ }
+ ],
+ "source": [
+ "prompt = PromptTemplate(template=\"{foo}{bar}\", input_variables=[\"foo\", \"bar\"])\n",
+ "partial_prompt = prompt.partial(foo=\"foo\");\n",
+ "print(partial_prompt.format(bar=\"baz\"))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "9986766e",
+ "metadata": {},
+ "source": [
+ "You can also just initialize the prompt with the partialed variables."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "id": "e2ce95b3",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "foobaz\n"
+ ]
+ }
+ ],
+ "source": [
+ "prompt = PromptTemplate(template=\"{foo}{bar}\", input_variables=[\"bar\"], partial_variables={\"foo\": \"foo\"})\n",
+ "print(prompt.format(bar=\"baz\"))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "a9c66f83",
+ "metadata": {},
+ "source": [
+ "## Partial With Functions\n",
+ "\n",
+ "The other common use is to partial with a function. The use case for this is when you have a variable you know that you always want to fetch in a common way. A prime example of this is with date or time. Imagine you have a prompt which you always want to have the current date. You can't hard code it in the prompt, and passing it along with the other input variables is a bit annoying. In this case, it's very handy to be able to partial the prompt with a function that always returns the current date."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "id": "d0712d8a",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from datetime import datetime\n",
+ "\n",
+ "def _get_datetime():\n",
+ " now = datetime.now()\n",
+ " return now.strftime(\"%m/%d/%Y, %H:%M:%S\")"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "id": "4cbcb666",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Tell me a funny joke about the day 02/27/2023, 22:15:16\n"
+ ]
+ }
+ ],
+ "source": [
+ "prompt = PromptTemplate(\n",
+ " template=\"Tell me a {adjective} joke about the day {date}\", \n",
+ " input_variables=[\"adjective\", \"date\"]\n",
+ ");\n",
+ "partial_prompt = prompt.partial(date=_get_datetime)\n",
+ "print(partial_prompt.format(adjective=\"funny\"))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "ffed6811",
+ "metadata": {},
+ "source": [
+ "You can also just initialize the prompt with the partialed variables, which often makes more sense in this workflow."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "id": "96285b25",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Tell me a funny joke about the day 02/27/2023, 22:15:16\n"
+ ]
+ }
+ ],
+ "source": [
+ "prompt = PromptTemplate(\n",
+ " template=\"Tell me a {adjective} joke about the day {date}\", \n",
+ " input_variables=[\"adjective\"],\n",
+ " partial_variables={\"date\": _get_datetime}\n",
+ ");\n",
+ "print(prompt.format(adjective=\"funny\"))"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "4bff16f7",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.9.1"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

docs/modules/prompts/how_to_guides.rst

@@ -17,6 +17,8 @@ The user guide here shows more advanced workflows and how to use the library in
 
 `Few Shot Prompt Examples <./examples/few_shot_examples.html>`_: Examples of Few Shot Prompt Templates.
 
+`Partial Prompt Template <./examples/partial.html>`_: How to partial Prompt Templates.
+
 
 
 .. toctree::

langchain/prompts/base.py

@@ -1,12 +1,14 @@
 """BasePrompt schema definition."""
+from __future__ import annotations
+
 import json
 import re
 from abc import ABC, abstractmethod
 from pathlib import Path
-from typing import Any, Callable, Dict, List, Optional, Union
+from typing import Any, Callable, Dict, List, Mapping, Optional, Union
 
 import yaml
-from pydantic import BaseModel, Extra, root_validator
+from pydantic import BaseModel, Extra, Field, root_validator
 
 from langchain.formatting import formatter
 
@@ -117,6 +119,9 @@ class BasePromptTemplate(BaseModel, ABC):
  """A list of the names of the variables the prompt template expects."""
  output_parser: Optional[BaseOutputParser] = None
  """How to parse the output of calling an LLM on this formatted prompt."""
+ partial_variables: Mapping[str, Union[str, Callable[[], str]]] = Field(
+ default_factory=dict
+ )
 
  class Config:
  """Configuration for this pydantic object."""
@@ -132,8 +137,38 @@ def validate_variable_names(cls, values: Dict) -> Dict:
  "Cannot have an input variable named 'stop', as it is used internally,"
  " please rename."
  )
+ if "stop" in values["partial_variables"]:
+ raise ValueError(
+ "Cannot have an partial variable named 'stop', as it is used "
+ "internally, please rename."
+ )
+
+ overall = set(values["input_variables"]).intersection(
+ values["partial_variables"]
+ )
+ if overall:
+ raise ValueError(
+ f"Found overlapping input and partial variables: {overall}"
+ )
  return values
 
+ def partial(self, **kwargs: Union[str, Callable[[], str]]) -> BasePromptTemplate:
+ """Return a partial of the prompt template."""
+ prompt_dict = self.__dict__.copy()
+ prompt_dict["input_variables"] = list(
+ set(self.input_variables).difference(kwargs)
+ )
+ prompt_dict["partial_variables"] = {**self.partial_variables, **kwargs}
+ return type(self)(**prompt_dict)
+
+ def _merge_partial_and_user_variables(self, **kwargs: Any) -> Dict[str, Any]:
+ # Get partial params:
+ partial_kwargs = {
+ k: v if isinstance(v, str) else v()
+ for k, v in self.partial_variables.items()
+ }
+ return {**partial_kwargs, **kwargs}
+
  @abstractmethod
  def format(self, **kwargs: Any) -> str:
  """Format the prompt with the inputs.
@@ -173,6 +208,8 @@ def save(self, file_path: Union[Path, str]) -> None:
 
  prompt.save(file_path="path/prompt.yaml")
  """
+ if self.partial_variables:
+ raise ValueError("Cannot save prompt with partial variables.")
  # Convert file to Path object.
  if isinstance(file_path, str):
  save_path = Path(file_path)

langchain/prompts/few_shot.py

@@ -68,7 +68,7 @@ def template_is_valid(cls, values: Dict) -> Dict:
  check_valid_template(
  values["prefix"] + values["suffix"],
  values["template_format"],
- values["input_variables"],
+ values["input_variables"] + list(values["partial_variables"]),
  )
  return values
 
@@ -101,6 +101,7 @@ def format(self, **kwargs: Any) -> str:
 
  prompt.format(variable1="foo")
  """
+ kwargs = self._merge_partial_and_user_variables(**kwargs)
  # Get the examples to use.
  examples = self._get_examples(**kwargs)
  # Format the examples.
@@ -110,6 +111,7 @@ def format(self, **kwargs: Any) -> str:
  # Create the overall template.
  pieces = [self.prefix, *example_strings, self.suffix]
  template = self.example_separator.join([piece for piece in pieces if piece])
+
  # Format the template with the input variables.
  return DEFAULT_FORMATTER_MAPPING[self.template_format](template, **kwargs)
 

langchain/prompts/few_shot_with_templates.py

@@ -60,16 +60,18 @@ def check_examples_and_selector(cls, values: Dict) -> Dict:
  @root_validator()
  def template_is_valid(cls, values: Dict) -> Dict:
  """Check that prefix, suffix and input variables are consistent."""
- input_variables = values["input_variables"]
- expected_input_variables = set(values["suffix"].input_variables)
- if values["prefix"] is not None:
- expected_input_variables |= set(values["prefix"].input_variables)
- missing_vars = expected_input_variables.difference(input_variables)
- if missing_vars:
- raise ValueError(
- f"Got input_variables={input_variables}, but based on prefix/suffix "
- f"expected {expected_input_variables}"
- )
+ if values["validate_template"]:
+ input_variables = values["input_variables"]
+ expected_input_variables = set(values["suffix"].input_variables)
+ expected_input_variables |= set(values["partial_variables"])
+ if values["prefix"] is not None:
+ expected_input_variables |= set(values["prefix"].input_variables)
+ missing_vars = expected_input_variables.difference(input_variables)
+ if missing_vars:
+ raise ValueError(
+ f"Got input_variables={input_variables}, but based on "
+ f"prefix/suffix expected {expected_input_variables}"
+ )
  return values
 
  class Config:
@@ -101,6 +103,7 @@ def format(self, **kwargs: Any) -> str:
 
  prompt.format(variable1="foo")
  """
+ kwargs = self._merge_partial_and_user_variables(**kwargs)
  # Get the examples to use.
  examples = self._get_examples(**kwargs)
  # Format the examples.

langchain/prompts/prompt.py

@@ -60,14 +60,16 @@ def format(self, **kwargs: Any) -> str:
 
  prompt.format(variable1="foo")
  """
+ kwargs = self._merge_partial_and_user_variables(**kwargs)
  return DEFAULT_FORMATTER_MAPPING[self.template_format](self.template, **kwargs)
 
  @root_validator()
  def template_is_valid(cls, values: Dict) -> Dict:
  """Check that template and input variables are consistent."""
  if values["validate_template"]:
+ all_inputs = values["input_variables"] + list(values["partial_variables"])
  check_valid_template(
- values["template"], values["template_format"], values["input_variables"]
+ values["template"], values["template_format"], all_inputs
  )
  return values