Skip to content

armgit / dspy

Public Imported from GitHub
0

DSPy: The framework for programming—not prompting—language models

Python

WIP for raw_demos in ChatAdapter

Omar Khattab2b32abd0d1f5
diff --git a/dspy/adapters/chat_adapter.py b/dspy/adapters/chat_adapter.py
index f7928ed17..627f8322a 100644
--- a/dspy/adapters/chat_adapter.py
+++ b/dspy/adapters/chat_adapter.py
@@ -1,7 +1,7 @@
import re
from .base import Adapter
-field_header_pattern = re.compile(r'\[\[\[\[ #### (\w+) #### \]\]\]\]')
+field_header_pattern = re.compile(r'\[\[\[ ### (\w+) ### \]\]\]')
class ChatAdapter(Adapter):
@@ -10,7 +10,15 @@ class ChatAdapter(Adapter):
def format(self, signature, demos, inputs):
messages = []
+
+ # TODO: Extract `raw_demos` out of `demos`, i.e. demos where some of the output_fields are not filled in.
+ # raw_demos = [demo for demo in demos if not all(k in demo for k in signature.output_fields)]
+ # demos = [demo for demo in demos if demo not in raw_demos]
+
messages.append({"role": "system", "content": prepare_instructions(signature)})
+ # messages.append({"role": "system", "content": prepare_instructions(signature, raw_demos)})
+
+ # TODO: Remove the raw_demos from demos.
for demo in demos:
output_fields_, demo_ = list(signature.output_fields.keys()) + ['completed'], {**demo, 'completed': ''}
@@ -42,13 +50,14 @@ class ChatAdapter(Adapter):
def format_fields(fields):
- return '\n\n'.join([f"[[[[ #### {k} #### ]]]]\n{v}" for k, v in fields.items()]).strip()
+ return '\n\n'.join([f"[[[ ### {k} ### ]]]\n{v}" for k, v in fields.items()]).strip()
def format_chat_turn(field_names, values):
- if not set(values).issuperset(set(field_names)):
- raise ValueError(f"Expected {field_names} but got {values.keys()}")
+ # TODO: Reinstate validation after dealing with raw_demos in the system messages.
+ # if not set(values).issuperset(set(field_names)):
+ # raise ValueError(f"Expected {field_names} but got {values.keys()}")
- return format_fields({k: values[k] for k in field_names})
+ return format_fields({k: values.get(k, "Not supplied for this particular example.") for k in field_names})
def enumerate_fields(fields):
parts = []
@@ -69,12 +78,12 @@ def prepare_instructions(signature):
parts.append(format_fields({f : f"{{{f}}}" for f in signature.output_fields}))
parts.append(format_fields({'completed' : ""}))
+ objective = ('\n' + ' ' * 8).join([''] + signature.instructions.splitlines())
+ parts.append(f"In adhering to this structure, your objective is: {objective}")
+
parts.append("You will receive some input fields in each interaction. " +
"Respond only with the corresponding output fields, starting with the field " +
", then ".join(f"`{f}`" for f in signature.output_fields) +
", and then ending with the marker for `completed`.")
-
- objective = ('\n' + ' ' * 8).join([''] + signature.instructions.splitlines())
- parts.append(f"In adhering to this structure, your objective is: {objective}")
return '\n\n'.join(parts).strip()
diff --git a/dspy/predict/chain_of_thought.py b/dspy/predict/chain_of_thought.py
index 16e1ba3e6..a7aa7b9a8 100644
--- a/dspy/predict/chain_of_thought.py
+++ b/dspy/predict/chain_of_thought.py
@@ -18,7 +18,11 @@ class ChainOfThought(Module):
*_keys, last_key = signature.output_fields.keys()
prefix = "Reasoning: Let's think step by step in order to"
- if dspy.settings.experimental:
+
+ import dspy
+ if isinstance(dspy.settings.lm, dspy.LM):
+ desc = "${reasoning}"
+ elif dspy.settings.experimental:
desc = "${produce the output fields}. We ..."
else:
desc = f"${{produce the {last_key}}}. We ..."