Skip to content

armgit / dspy

Public Imported from GitHub
0

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

Python

added configurable backoff max time

Arnav Singhvif4d4b7368d02
diff --git a/dsp/modules/anthropic.py b/dsp/modules/anthropic.py
index c343e2a2a..013988281 100644
--- a/dsp/modules/anthropic.py
+++ b/dsp/modules/anthropic.py
@@ -3,6 +3,7 @@ import os
from typing import Any, Optional
import backoff
+import dspy
from dsp.modules.lm import LM
@@ -90,7 +91,7 @@ class Claude(LM):
@backoff.on_exception(
backoff.expo,
(anthropic_rate_limit),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
max_tries=8,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
diff --git a/dsp/modules/aws_providers.py b/dsp/modules/aws_providers.py
index 5a6138ead..e350c8898 100644
--- a/dsp/modules/aws_providers.py
+++ b/dsp/modules/aws_providers.py
@@ -4,6 +4,7 @@ from abc import ABC, abstractmethod
from typing import Any, Optional
import backoff
+import dspy
try:
import boto3
@@ -80,7 +81,7 @@ class AWSProvider(ABC):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
max_tries=8,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
@@ -155,7 +156,7 @@ class Sagemaker(AWSProvider):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
max_tries=8,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
diff --git a/dsp/modules/azure_openai.py b/dsp/modules/azure_openai.py
index bc4fa61ce..45ca89bfd 100644
--- a/dsp/modules/azure_openai.py
+++ b/dsp/modules/azure_openai.py
@@ -8,6 +8,7 @@ import openai
from dsp.modules.cache_utils import CacheMemory, NotebookCacheMemory, cache_turn_on
from dsp.modules.lm import LM
+import dspy
try:
OPENAI_LEGACY = int(openai.version.__version__[0]) == 0
@@ -166,7 +167,7 @@ class AzureOpenAI(LM):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
)
def request(self, prompt: str, **kwargs):
diff --git a/dsp/modules/cloudflare.py b/dsp/modules/cloudflare.py
index 81c20f838..abb3e5179 100644
--- a/dsp/modules/cloudflare.py
+++ b/dsp/modules/cloudflare.py
@@ -7,6 +7,7 @@ import requests
from pydantic import BaseModel, ValidationError
from dsp.modules.lm import LM
+import dspy
def backoff_hdlr(details) -> None:
@@ -69,7 +70,7 @@ class CloudflareAI(LM):
@backoff.on_exception(
backoff.expo,
requests.exceptions.RequestException,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
on_giveup=giveup_hdlr,
)
diff --git a/dsp/modules/cohere.py b/dsp/modules/cohere.py
index 401fe9546..67882b027 100644
--- a/dsp/modules/cohere.py
+++ b/dsp/modules/cohere.py
@@ -3,6 +3,7 @@ from typing import Any, Optional
import backoff
from dsp.modules.lm import LM
+import dspy
try:
import cohere
@@ -101,7 +102,7 @@ class Cohere(LM):
@backoff.on_exception(
backoff.expo,
(cohere_api_error),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
)
diff --git a/dsp/modules/google.py b/dsp/modules/google.py
index 5d97534ea..07e9922e7 100644
--- a/dsp/modules/google.py
+++ b/dsp/modules/google.py
@@ -5,6 +5,7 @@ from typing import Any, Optional
import backoff
from dsp.modules.lm import LM
+import dspy
try:
import google.generativeai as genai
@@ -133,7 +134,7 @@ class Google(LM):
@backoff.on_exception(
backoff.expo,
(google_api_error),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
max_tries=8,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
diff --git a/dsp/modules/googlevertexai.py b/dsp/modules/googlevertexai.py
index 118480040..5db14ebe3 100644
--- a/dsp/modules/googlevertexai.py
+++ b/dsp/modules/googlevertexai.py
@@ -6,6 +6,7 @@ import backoff
from pydantic_core import PydanticCustomError
from dsp.modules.lm import LM
+import dspy
try:
import vertexai # type: ignore[import-untyped]
@@ -184,7 +185,7 @@ class GoogleVertexAI(LM):
@backoff.on_exception(
backoff.expo,
(Exception),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
)
diff --git a/dsp/modules/gpt3.py b/dsp/modules/gpt3.py
index 54c0af012..c779e4359 100644
--- a/dsp/modules/gpt3.py
+++ b/dsp/modules/gpt3.py
@@ -8,6 +8,7 @@ import openai
from dsp.modules.cache_utils import CacheMemory, NotebookCacheMemory, cache_turn_on
from dsp.modules.lm import LM
+import dspy
try:
OPENAI_LEGACY = int(openai.version.__version__[0]) == 0
@@ -133,11 +134,11 @@ class GPT3(LM):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
)
def request(self, prompt: str, **kwargs):
- """Handles retreival of GPT-3 completions whilst handling rate limiting and caching."""
+ """Handles retrieval of GPT-3 completions whilst handling rate limiting and caching."""
if "model_type" in kwargs:
del kwargs["model_type"]
diff --git a/dsp/modules/groq_client.py b/dsp/modules/groq_client.py
index 6650897ae..c62e3a3cd 100644
--- a/dsp/modules/groq_client.py
+++ b/dsp/modules/groq_client.py
@@ -13,6 +13,7 @@ except ImportError:
from dsp.modules.lm import LM
+import dspy
def backoff_hdlr(details):
@@ -91,7 +92,7 @@ class GroqLM(LM):
@backoff.on_exception(
backoff.expo,
groq_api_error,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
)
def request(self, prompt: str, **kwargs):
diff --git a/dsp/modules/hf_client.py b/dsp/modules/hf_client.py
index 368f96c2b..f11a613e9 100644
--- a/dsp/modules/hf_client.py
+++ b/dsp/modules/hf_client.py
@@ -11,6 +11,7 @@ import requests
from dsp.modules.cache_utils import CacheMemory, NotebookCacheMemory
from dsp.modules.hf import HFModel, openai_to_hf
+import dspy
ERRORS = (Exception)
@@ -340,7 +341,7 @@ class Together(HFModel):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
)
def _generate(self, prompt, use_chat_api=False, **kwargs):
diff --git a/dsp/modules/mistral.py b/dsp/modules/mistral.py
index abe5d9a23..ecb79070e 100644
--- a/dsp/modules/mistral.py
+++ b/dsp/modules/mistral.py
@@ -3,6 +3,7 @@ from typing import Any, Optional
import backoff
from dsp.modules.lm import LM
+import dspy
try:
import mistralai
@@ -99,7 +100,7 @@ class Mistral(LM):
@backoff.on_exception(
backoff.expo,
(mistralai_api_error),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
)
diff --git a/dsp/modules/premai.py b/dsp/modules/premai.py
index 80345824a..166fc72d5 100644
--- a/dsp/modules/premai.py
+++ b/dsp/modules/premai.py
@@ -5,6 +5,7 @@ from typing import Any, Optional
import backoff
from dsp.modules.lm import LM
+import dspy
try:
import premai
@@ -179,7 +180,7 @@ class PremAI(LM):
@backoff.on_exception(
backoff.expo,
(premai_api_error),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
)
diff --git a/dsp/modules/snowflake.py b/dsp/modules/snowflake.py
index 09bffd787..ad308abc0 100644
--- a/dsp/modules/snowflake.py
+++ b/dsp/modules/snowflake.py
@@ -6,6 +6,7 @@ import backoff
from pydantic_core import PydanticCustomError
from dsp.modules.lm import LM
+import dspy
try:
from snowflake.snowpark import Session
@@ -146,7 +147,7 @@ class Snowflake(LM):
@backoff.on_exception(
backoff.expo,
(Exception),
- max_time=1000,
+ max_time=dspy.settings.backoff_time,
on_backoff=backoff_hdlr,
giveup=giveup_hdlr,
)
diff --git a/dsp/utils/settings.py b/dsp/utils/settings.py
index 37cab1264..e7fe58cbd 100644
--- a/dsp/utils/settings.py
+++ b/dsp/utils/settings.py
@@ -42,6 +42,7 @@ class Settings:
suggest_failures=0,
langchain_history=[],
experimental=False,
+ backoff_time = 10
)
cls._instance.__append(config)
diff --git a/dspy/retrieve/chromadb_rm.py b/dspy/retrieve/chromadb_rm.py
index aa8589d75..97a9bba8d 100644
--- a/dspy/retrieve/chromadb_rm.py
+++ b/dspy/retrieve/chromadb_rm.py
@@ -124,7 +124,7 @@ class ChromadbRM(dspy.Retrieve):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=15,
+ max_time=dspy.settings.backoff_time,
)
def _get_embeddings(self, queries: List[str]) -> List[List[float]]:
"""Return query vector after creating embedding using OpenAI
diff --git a/dspy/retrieve/mongodb_atlas_rm.py b/dspy/retrieve/mongodb_atlas_rm.py
index 3cafb05f6..8b2c76c54 100644
--- a/dspy/retrieve/mongodb_atlas_rm.py
+++ b/dspy/retrieve/mongodb_atlas_rm.py
@@ -61,7 +61,7 @@ class Embedder:
RateLimitError,
UnprocessableEntityError,
),
- max_time=15,
+ max_time=dspy.settings.backoff_time,
)
def __call__(self, queries) -> Any:
embedding = self.client.embeddings.create(input=queries, model=self.model)
diff --git a/dspy/retrieve/neo4j_rm.py b/dspy/retrieve/neo4j_rm.py
index bb5ad2523..d8c071277 100644
--- a/dspy/retrieve/neo4j_rm.py
+++ b/dspy/retrieve/neo4j_rm.py
@@ -42,7 +42,7 @@ class Embedder:
RateLimitError,
UnprocessableEntityError,
),
- max_time=15,
+ max_time=dspy.settings.backoff_time,
)
def __call__(self, queries) -> Any:
embedding = self.client.embeddings.create(input=queries, model=self.model)
diff --git a/dspy/retrieve/pinecone_rm.py b/dspy/retrieve/pinecone_rm.py
index 0328bc6ba..5d527f8dd 100644
--- a/dspy/retrieve/pinecone_rm.py
+++ b/dspy/retrieve/pinecone_rm.py
@@ -178,7 +178,7 @@ class PineconeRM(dspy.Retrieve):
@backoff.on_exception(
backoff.expo,
ERRORS,
- max_time=15,
+ max_time=dspy.settings.backoff_time,
)
def _get_embeddings(
self,