updated examples

Author: qhoxie

README.md

@@ -399,7 +399,7 @@ extractor = LangStruct(
     schema=YourSchema,
     model="gemini/gemini-2.5-flash",  # Fast & cost-effective
     chunking_config=config,
-    optimize=True  # Default is True
+    optimize=True  # Enabled for training data
 )
 ```
 

docs/src/content/docs/examples/financial-documents.mdx

@@ -117,8 +117,8 @@ Extract key information from quarterly earnings reports:
     # Create specialized financial document extractor
     financial_extractor = LangStruct(
         schema=FinancialReportSchema,
-        model="gemini/gemini-2.5-flash"  # Fast and cost-effective for financial data
-        # Auto-optimization and source grounding enabled by default
+        model="gemini/gemini-2.5-flash-lite"  # Fast and cost-effective for financial data
+        # Source grounding enabled by default, optimization can be enabled if needed
     )
 
     # Sample earnings report text

docs/src/content/docs/examples/legal-contracts.mdx

@@ -64,7 +64,7 @@ Create an extractor for legal document analysis:
 # Create the extractor with legal domain optimization
 extractor = LangStruct(
     schema=LegalContractSchema,
-    model="gemini/gemini-2.5-flash",  # Fast and reliable for legal analysis
+    model="gemini/gemini-2.5-flash-lite",  # Fast and reliable for legal analysis
     optimize=True,
     use_sources=True,  # Critical for legal document traceability
     temperature=0.1,   # Lower temperature for consistency
@@ -190,7 +190,7 @@ class ContractRiskAssessment(Schema):
 
 risk_analyzer = LangStruct(
     schema=ContractRiskAssessment,
-    model="gemini/gemini-2.5-flash",
+    model="gemini/gemini-2.5-flash-lite",
     system_prompt="""You are a legal risk analyst. Identify potential risks in contracts 
     and provide actionable recommendations for risk mitigation. Focus on:
     - Financial exposure and liability
@@ -412,7 +412,7 @@ You are analyzing a Non-Disclosure Agreement (NDA). Focus on:
 nda_extractor = LangStruct(
     schema=NDASchema,
     system_prompt=nda_prompt,
-    model="gemini/gemini-2.5-flash"
+    model="gemini/gemini-2.5-flash-lite"
 )
 ```
 

docs/src/content/docs/examples/medical-records.mdx

@@ -91,7 +91,7 @@ Extract key information from clinical notes:
 # Create medical data extractor
 medical_extractor = LangStruct(
     schema=MedicalRecordSchema,
-    model="gemini/gemini-2.5-flash",  # Fast and reliable for medical analysis
+    model="gemini/gemini-2.5-flash-lite",  # Fast and reliable for medical analysis
     temperature=0.0,  # Zero temperature for consistent medical analysis
     use_sources=True  # Track sources for validation
 )

docs/src/content/docs/examples/scientific-papers.mdx

@@ -74,7 +74,7 @@ Create an extractor for research paper analysis:
 # Create the extractor optimized for academic content
 extractor = LangStruct(
     schema=ScientificPaperSchema,
-    model="gemini/gemini-2.5-flash",  # Fast and reliable for academic content
+    model="gemini/gemini-2.5-flash-lite",  # Fast and reliable for academic content
     optimize=True,
     use_sources=True,  # Track where information was found
     temperature=0.2,   # Slightly higher for nuanced interpretation
@@ -504,7 +504,7 @@ You are analyzing a medical research paper. Focus on:
 medical_extractor = LangStruct(
     schema=MedicalPaperSchema,
     system_prompt=medical_prompt,
-    model="gemini/gemini-2.5-flash"
+    model="gemini/gemini-2.5-flash-lite"
 )
 ```
 

docs/src/content/docs/installation.mdx

@@ -111,7 +111,7 @@ Set up any provider you prefer:
 export GOOGLE_API_KEY="your-google-api-key"
 ```
 
-Use any Gemini model: `gemini-2.5-flash`, `gemini-2.5-pro`, etc.
+Use any Gemini model: `gemini-2.5-flash-lite`, `gemini-2.5-flash`, `gemini-2.5-pro`, etc.
 
 ### OpenAI
 

docs/src/content/docs/optimization.mdx

@@ -9,32 +9,31 @@ Make your extraction more accurate with automatic optimization. LangStruct learn
 
 ## The Easy Way
 
-**Optimization is enabled by default - you're already getting better results:**
+**Enable optimization for better results:**
 
 ```python
 from langstruct import LangStruct
 
-# Create extractor (optimization enabled by default)
-extractor = LangStruct(example={
-    "name": "Dr. Sarah Johnson",
-    "age": 34,
-    "occupation": "data scientist"
-})
+# Create extractor with optimization enabled
+extractor = LangStruct(
+    example={
+        "name": "Dr. Sarah Johnson",
+        "age": 34,
+        "occupation": "data scientist"
+    },
+    optimize=True
+)
 
 result = extractor.extract("Dr. Sarah Johnson, 34, is a data scientist")
-print(result.entities)  # Already optimized results!
+print(result.entities)  # Optimized results!
 ```
 
-That's it! Your extractions are automatically improving over time.
-
-**To disable optimization (not recommended):**
+**Default behavior (faster startup, good baseline accuracy):**
 
 ```python
-# Only if you need faster startup and don't care about accuracy
-extractor = LangStruct(
-    example={"name": "John", "age": 30}, 
-    optimize=False
-)
+# No optimization - good for quick experiments
+extractor = LangStruct(example={"name": "John", "age": 30})
+# optimize=False by default - enables faster startup
 ```
 
 ## When You Have Training Data
@@ -106,7 +105,7 @@ extractor.optimize(
 
 <CardGrid>
 	<Card title="Start Simple" icon="rocket">
-		Optimization is enabled by default - just create your extractor
+		Start without optimization for quick experiments, enable when you need accuracy
 	</Card>
 	<Card title="Quality Over Quantity" icon="star">
 		10 good training examples beats 100 poor ones
@@ -122,7 +121,7 @@ extractor.optimize(
 ## Common Questions
 
 **Q: Do I always need training data?**  
-A: No! Optimization works without any training data and still improves results.
+A: No! Optimization can work without training data, but providing examples improves results significantly.
 
 **Q: How long does optimization take?**  
 A: Usually 1-5 minutes for typical datasets (10-100 examples).
@@ -140,7 +139,7 @@ A: Just change the model and re-optimize! Same training data, same accuracy - ze
 
 <CardGrid>
 	<Card title="Try It Now" icon="laptop">
-		Create a LangStruct extractor - optimization is already enabled!
+		Create a LangStruct extractor and enable optimization when you need accuracy!
 	</Card>
 	<Card title="Source Grounding" icon="document">
 		[Track where information comes from](/source-grounding/)

docs/src/content/docs/quickstart.mdx

@@ -166,7 +166,7 @@ extractor = LangStruct(example=schema)
 # Or specify model explicitly
 extractor = LangStruct(
     example=schema,
-    model="gemini/gemini-2.5-flash"  # Fast & cheap
+    model="gemini/gemini-2.5-flash-lite"  # Fast & cheap
 )
 
 # Local models

docs/src/content/docs/source-grounding.mdx

@@ -44,7 +44,7 @@ class PersonSchema(Schema):
 # Enable source grounding
 extractor = LangStruct(
     schema=PersonSchema,
-    model="gemini/gemini-2.5-flash",  # Fast and cost-effective
+    model="gemini/gemini-2.5-flash-lite",  # Fast and cost-effective
     use_sources=True  # Enable source tracking
 )
 
@@ -323,7 +323,7 @@ class ComplianceExtractor:
         self.extractor = LangStruct(
             schema=schema,
             use_sources=True,
-            model="gemini/gemini-2.5-flash"  # Fast and reliable
+            model="gemini/gemini-2.5-flash-lite"  # Fast and reliable
         )
 
     def extract_with_audit_trail(self, document, document_id=None):

docs/src/content/docs/why-dspy.mdx

@@ -146,7 +146,8 @@ result = extractor.extract("Microsoft announced $65B revenue for Q4")
 # Start with OpenAI
 extractor = LangStruct(
     example={"company": "Apple", "revenue": 100.0}, 
-    model="gpt-4o"
+    model="gpt-4o",
+    optimize=True
 )
 extractor.optimize(training_texts, expected_results)