SimplerLLM 0.3.1.10__tar.gz → 0.3.1.11__tar.gz

{simplerllm-0.3.1.10 → simplerllm-0.3.1.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: SimplerLLM
-Version: 0.3.1.10
+Version: 0.3.1.11
 Summary: An easy-to-use Library for interacting with language models.
 Home-page: https://github.com/hassancs91/SimplerLLM
 Author: Hasan Aboul Hasan

{simplerllm-0.3.1.10 → simplerllm-0.3.1.11}/SimplerLLM/language/llm_addons.py

@@ -124,7 +124,7 @@ def generate_pydantic_json_model_reliable(
     custom_prompt_suffix: str = None,
     system_prompt: str = "The Output is a VALID Structured JSON",
     full_response: bool = False,
-) -> Union[Tuple[BaseModel, LLMProvider], Tuple[BaseModel, LLMFullResponse, LLMProvider]]:
+) -> Union[Tuple[BaseModel, LLMProvider, str], Tuple[BaseModel, LLMFullResponse, LLMProvider, str]]:
     """
     Generates a model instance using ReliableLLM with fallback capability.
 
@@ -132,10 +132,16 @@ def generate_pydantic_json_model_reliable(
     :param prompt: The fully formatted prompt including the topic.
     :param reliable_llm: Instance of ReliableLLM with primary and secondary providers.
     :param max_retries: Maximum number of retries on validation errors.
+    :param max_tokens: Maximum number of tokens to generate.
     :param initial_delay: Initial delay in seconds before the first retry.
     :param custom_prompt_suffix: Optional string to customize or override the generated prompt extension.
+    :param system_prompt: System prompt to set the context for the LLM.
+    :param full_response: If True, returns the full API response.
 
-    :return: BaseModel object if successful, otherwise error message.
+    :return: 
+        - If full_response=False: Tuple of (model_object, provider, model_name)
+        - If full_response=True: Tuple of (model_object, full_response, provider, model_name)
+        - Error message string if unsuccessful
     """
     optimized_prompt = create_optimized_prompt(prompt, model_class, custom_prompt_suffix)
     backoff_delays = [initial_delay * (2**attempt) for attempt in range(max_retries + 1)]
@@ -152,10 +158,10 @@ def generate_pydantic_json_model_reliable(
             )
             
             if full_response:
-                ai_response, provider = result
+                ai_response, provider, model_name = result
                 response_text = ai_response.generated_text
             else:
-                response_text, provider = result
+                response_text, provider, model_name = result
 
             if response_text:
                 json_object = extract_json_from_text(response_text)
@@ -168,7 +174,7 @@ def generate_pydantic_json_model_reliable(
                     model_object = convert_json_to_pydantic_model(
                         model_class, json_object[0]
                     )
-                    return (model_object, ai_response, provider) if full_response else (model_object, provider)
+                    return (model_object, ai_response, provider, model_name) if full_response else (model_object, provider, model_name)
 
         except Exception as e:
             return f"Exception occurred: {e}"
@@ -190,7 +196,7 @@ async def generate_pydantic_json_model_reliable_async(
     custom_prompt_suffix: str = None,
     system_prompt: str = "The Output is a VALID Structured JSON",
     full_response: bool = False,
-) -> Union[Tuple[BaseModel, LLMProvider], Tuple[BaseModel, LLMFullResponse, LLMProvider]]:
+) -> Union[Tuple[BaseModel, LLMProvider, str], Tuple[BaseModel, LLMFullResponse, LLMProvider, str]]:
     """
     Asynchronously generates a model instance using ReliableLLM with fallback capability.
 
@@ -198,10 +204,16 @@ async def generate_pydantic_json_model_reliable_async(
     :param prompt: The fully formatted prompt including the topic.
     :param reliable_llm: Instance of ReliableLLM with primary and secondary providers.
     :param max_retries: Maximum number of retries on validation errors.
+    :param max_tokens: Maximum number of tokens to generate.
     :param initial_delay: Initial delay in seconds before the first retry.
     :param custom_prompt_suffix: Optional string to customize or override the generated prompt extension.
+    :param system_prompt: System prompt to set the context for the LLM.
+    :param full_response: If True, returns the full API response.
 
-    :return: BaseModel object if successful, otherwise error message.
+    :return: 
+        - If full_response=False: Tuple of (model_object, provider, model_name)
+        - If full_response=True: Tuple of (model_object, full_response, provider, model_name)
+        - Error message string if unsuccessful
     """
     optimized_prompt = create_optimized_prompt(prompt, model_class, custom_prompt_suffix)
     backoff_delays = [initial_delay * (2**attempt) for attempt in range(max_retries + 1)]
@@ -218,10 +230,10 @@ async def generate_pydantic_json_model_reliable_async(
             )
             
             if full_response:
-                ai_response, provider = result
+                ai_response, provider, model_name = result
                 response_text = ai_response.generated_text
             else:
-                response_text, provider = result
+                response_text, provider, model_name = result
 
             if response_text:
                 json_object = extract_json_from_text(response_text)
@@ -234,7 +246,7 @@ async def generate_pydantic_json_model_reliable_async(
                     model_object = convert_json_to_pydantic_model(
                         model_class, json_object[0]
                     )
-                    return (model_object, ai_response, provider) if full_response else (model_object, provider)
+                    return (model_object, ai_response, provider, model_name) if full_response else (model_object, provider, model_name)
 
         except Exception as e:
             return f"Exception occurred: {e}"

{simplerllm-0.3.1.10 → simplerllm-0.3.1.11}/SimplerLLM.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: SimplerLLM
-Version: 0.3.1.10
+Version: 0.3.1.11
 Summary: An easy-to-use Library for interacting with language models.
 Home-page: https://github.com/hassancs91/SimplerLLM
 Author: Hasan Aboul Hasan

{simplerllm-0.3.1.10 → simplerllm-0.3.1.11}/SimplerLLM.egg-info/SOURCES.txt

@@ -91,7 +91,6 @@ tests/test_reAct_agent.py
 tests/test_react_core_agent.py
 tests/test_reflection_core_agent.py
 tests/test_reliable.py
-tests/test_reliable_json.py
 tests/test_search_query.py
 tests/test_sql_agent.py
 tests/test_token_count.py

{simplerllm-0.3.1.10 → simplerllm-0.3.1.11}/setup.py

@@ -11,7 +11,7 @@ with open("README.md", encoding="utf-8") as f:
 
 setup(
     name="SimplerLLM",
-    version="0.3.1.10",
+    version="0.3.1.11",
     author="Hasan Aboul Hasan",
     author_email="hasan@learnwithhasan.com",
     description="An easy-to-use Library for interacting with language models.",

simplerllm-0.3.1.10/tests/test_reliable_json.py

@@ -1,118 +0,0 @@
-from typing import List
-from pydantic import BaseModel
-from SimplerLLM.language.llm import LLM, LLMProvider
-from SimplerLLM.language.llm.reliable import ReliableLLM
-from SimplerLLM.language.llm_addons import generate_pydantic_json_model_reliable
-
-# Using the same model from test_complex_json.py for consistency
-class EffectivenessScore(BaseModel):
-    Score: int
-    Explanation: str
-
-class ScannabilityScore(BaseModel):
-    Score: int
-    Explanation: str
-
-class SentimentAnalysis(BaseModel):
-    Tone: str
-    Explanation: str
-
-class LengthAnalysis(BaseModel):
-    number_of_chars: str
-    Explanation: str
-
-class SpamTriggers(BaseModel):
-    Triggers: List[str]
-    Explanation: str
-
-class AllCapsWords(BaseModel):
-    Words: List[str]
-    Impact: str
-
-class Emojis(BaseModel):
-    Recommendation: str
-    Explanation: str
-
-class EmailSubjectLineAnalysis(BaseModel):
-    Effectiveness_Score: EffectivenessScore
-    Scannability_Score: ScannabilityScore
-    Sentiment_Analysis: SentimentAnalysis
-    Length_Analysis: LengthAnalysis
-    Spam_Triggers: SpamTriggers
-    All_Caps_Words: AllCapsWords
-    Emojis: Emojis
-    Suggested_Preview_Text: List[str]
-    Alternative_Subject_Lines: List[str]
-
-# Create primary and secondary LLM instances
-primary_llm = LLM.create(
-    provider=LLMProvider.ANTHROPIC,
-    model_name="claude-3-opus-latest",
-    verbose=True
-)
-
-secondary_llm = LLM.create(
-    provider=LLMProvider.OPENAI,
-    model_name="gpt-4o",
-)
-
-# Create ReliableLLM instance
-reliable_llm = ReliableLLM(primary_llm=primary_llm, secondary_llm=secondary_llm, verbose=True)
-
-# Test prompt
-subject_line_grader_prompt = """
-You are a professional email subject line copywriter tasked with analyzing and providing feedback on a given subject line. Your analysis should be thorough, considering multiple factors to ensure the best possible subject line optimization.
-
-Here is the information you need to analyze:
-
-<subject_line>
-{subject_line}
-</subject_line>
-
-<email_type>
-{email_type}
-</email_type>
-
-<is_time_sensitive>
-{is_time_sensitive}
-</is_time_sensitive>
-
-<is_followup>
-{is_followup}
-</is_followup>
-
-<is_series>
-{is_series}
-</is_series>
-
-Please analyze the subject line considering the following aspects:
-1. Effectiveness
-2. Scannability
-3. Sentiment
-4. Length
-5. Spam Triggers
-6. All Caps Words
-7. Emoji Usage
-"""
-
-# Format the prompt with test data
-generated_prompt = subject_line_grader_prompt.format(
-    subject_line="3 seo tips",
-    email_type="Newsletter",
-    is_time_sensitive="No",
-    is_followup="No",
-    is_series="No"
-)
-
-# Generate model using ReliableLLM
-model_response, used_provider = generate_pydantic_json_model_reliable(
-    model_class=EmailSubjectLineAnalysis,
-    prompt=generated_prompt,
-    reliable_llm=reliable_llm,
-    max_retries=2,
-    max_tokens=4096,
-    system_prompt="Generate in English Language"
-)
-
-print("Response from ReliableLLM using provider:", used_provider)
-print(model_response)