featrixsphere 0.2.193__tar.gz → 0.2.4984__tar.gz

featrixsphere-0.2.4984/MANIFEST.in

@@ -0,0 +1,8 @@
+include VERSION
+include README.md
+include requirements.txt
+include system_monitor.py
+include nv-install.sh
+include featrix-update.py
+recursive-include featrixsphere *.py
+recursive-include src *.py *.sh 

{featrixsphere-0.2.193/featrixsphere.egg-info → featrixsphere-0.2.4984}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: featrixsphere
-Version: 0.2.193
+Version: 0.2.4984
 Summary: Transform any CSV into a production-ready ML model in minutes, not months.
 Home-page: https://github.com/Featrix/sphere
 Author: Featrix
@@ -299,6 +299,47 @@ print(f"""
 
 ## 🔍 Advanced Features
 
+### Batch Encoding (NEW in v0.2.228! 🆕)
+```python
+# Encode single record
+embedding = client.encode_records(session_id, {
+    "text": "customer complaint about billing",
+    "category": "support"
+})
+
+# NEW: Batch encoding with intelligent adaptive sizing!
+# Handles any size - even 50,000+ records efficiently
+records = [
+    {"text": "happy customer", "category": "support"},
+    {"text": "billing issue", "category": "finance"},
+    # ... thousands more records
+]
+
+# Automatically batches, measures response time, and optimizes throughput
+embeddings = client.encode_records(session_id, records)
+# Output:
+# 📊 Encoding 50,000 records with adaptive batching...
+#   ✓ Batch: 100 records in 2.3s (43.5 rec/s) - Progress: 100/50,000 (0.2%)
+#   ⚡ Fast response, increasing batch size to 150
+#   ✓ Batch: 150 records in 3.1s (48.4 rec/s) - Progress: 250/50,000 (0.5%)
+#   ...
+# ✅ Completed encoding 50,000 records
+
+# Access both 3D and full-dimensional embeddings
+for result in embeddings:
+    short_3d = result['embedding_short']      # 3D for visualization
+    full_embedding = result['embedding_long']  # Full-dimensional for ML
+```
+
+**How Batch Encoding Works:**
+- Starts with batches of 100 records
+- Measures each batch response time
+- If < 3.5s → increases batch size by 1.5x
+- If < 5s → increases by 1.2x  
+- If > 6.5s → decreases by 0.7x
+- Targets ~5 seconds per batch for optimal throughput
+- Range: 10-5,000 records per batch
+
 ### Similarity Search
 ```python
 # Find similar records using neural embeddings
@@ -321,7 +362,8 @@ embedding = client.encode_records(session_id, {
     "priority": "high"
 })
 
-print(f"Embedding dimension: {len(embedding['embedding'])}")
+print(f"3D embedding: {embedding['embedding_short']}")
+print(f"Full embedding dimension: {len(embedding['embedding_long'])}")
 # Embedding dimension: 512  (rich 512-dimensional representation!)
 ```
 
@@ -386,14 +428,11 @@ print(f"""
 batch_results = client.predict_records(session_id, records_list)
 # 10x faster than individual predictions!
 
-# Adjust training parameters for your data size
+# Featrix will automatically tune your model for your data.
 client.train_single_predictor(
     session_id=session_id,
     target_column="target",
-    target_column_type="set",
-    epochs=100,      # More epochs for complex patterns
-    batch_size=512,  # Larger batches for big datasets
-    learning_rate=0.001  # Lower LR for stable training
+    target_column_type="set"
 )
 ```
 

{featrixsphere-0.2.193 → featrixsphere-0.2.4984}/README.md

@@ -219,6 +219,47 @@ print(f"""
 
 ## 🔍 Advanced Features
 
+### Batch Encoding (NEW in v0.2.228! 🆕)
+```python
+# Encode single record
+embedding = client.encode_records(session_id, {
+    "text": "customer complaint about billing",
+    "category": "support"
+})
+
+# NEW: Batch encoding with intelligent adaptive sizing!
+# Handles any size - even 50,000+ records efficiently
+records = [
+    {"text": "happy customer", "category": "support"},
+    {"text": "billing issue", "category": "finance"},
+    # ... thousands more records
+]
+
+# Automatically batches, measures response time, and optimizes throughput
+embeddings = client.encode_records(session_id, records)
+# Output:
+# 📊 Encoding 50,000 records with adaptive batching...
+#   ✓ Batch: 100 records in 2.3s (43.5 rec/s) - Progress: 100/50,000 (0.2%)
+#   ⚡ Fast response, increasing batch size to 150
+#   ✓ Batch: 150 records in 3.1s (48.4 rec/s) - Progress: 250/50,000 (0.5%)
+#   ...
+# ✅ Completed encoding 50,000 records
+
+# Access both 3D and full-dimensional embeddings
+for result in embeddings:
+    short_3d = result['embedding_short']      # 3D for visualization
+    full_embedding = result['embedding_long']  # Full-dimensional for ML
+```
+
+**How Batch Encoding Works:**
+- Starts with batches of 100 records
+- Measures each batch response time
+- If < 3.5s → increases batch size by 1.5x
+- If < 5s → increases by 1.2x  
+- If > 6.5s → decreases by 0.7x
+- Targets ~5 seconds per batch for optimal throughput
+- Range: 10-5,000 records per batch
+
 ### Similarity Search
 ```python
 # Find similar records using neural embeddings
@@ -241,7 +282,8 @@ embedding = client.encode_records(session_id, {
     "priority": "high"
 })
 
-print(f"Embedding dimension: {len(embedding['embedding'])}")
+print(f"3D embedding: {embedding['embedding_short']}")
+print(f"Full embedding dimension: {len(embedding['embedding_long'])}")
 # Embedding dimension: 512  (rich 512-dimensional representation!)
 ```
 
@@ -306,14 +348,11 @@ print(f"""
 batch_results = client.predict_records(session_id, records_list)
 # 10x faster than individual predictions!
 
-# Adjust training parameters for your data size
+# Featrix will automatically tune your model for your data.
 client.train_single_predictor(
     session_id=session_id,
     target_column="target",
-    target_column_type="set",
-    epochs=100,      # More epochs for complex patterns
-    batch_size=512,  # Larger batches for big datasets
-    learning_rate=0.001  # Lower LR for stable training
+    target_column_type="set"
 )
 ```
 

featrixsphere-0.2.4984/VERSION

@@ -0,0 +1 @@
+0.2.4984