schema-search 0.1.4__tar.gz → 0.1.6__tar.gz

{schema_search-0.1.4/schema_search.egg-info → schema_search-0.1.6}/PKG-INFO

@@ -1,18 +1,16 @@
 Metadata-Version: 2.4
 Name: schema-search
-Version: 0.1.4
-Summary: Natural language search for database schemas with graph-aware semantic retrieval
-Home-page: https://github.com/neehan/schema-search
-Author: 
+Version: 0.1.6
+Summary: Natural language database schema search with graph-aware semantic retrieval
+Home-page: https://adibhasan.com/blog/schema-search/
+Author: Adib Hasan
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.8
-Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
-Requires-Python: >=3.8
+Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: sqlalchemy>=1.4.0
@@ -40,6 +38,7 @@ Requires-Dist: snowflake-sqlalchemy>=1.4.0; extra == "snowflake"
 Requires-Dist: snowflake-connector-python>=3.0.0; extra == "snowflake"
 Provides-Extra: bigquery
 Requires-Dist: sqlalchemy-bigquery>=1.6.0; extra == "bigquery"
+Dynamic: author
 Dynamic: classifier
 Dynamic: description
 Dynamic: description-content-type
@@ -211,14 +210,14 @@ We [benchmarked](/tests/test_spider_eval.py) on the Spider dataset (1,234 train
 **Memory:** The embedding model requires ~90 MB and the optional reranker adds ~155 MB. Actual process memory depends on your Python runtime.
 
 ### Without Reranker (`reranker.model: null`)
-![Without Reranker](img/spider_benchmark_without_reranker.png)
+![Without Reranker](https://raw.githubusercontent.com/Neehan/schema-search/refs/heads/main/img/spider_benchmark_without_reranker.png)
 - **Indexing:** 0.22s ± 0.08s per database (18 total).
 - **Accuracy:** Hybrid leads with Recall@1 62% / MRR 0.93; Semantic follows at Recall@1 58% / MRR 0.89.
 - **Latency:** BM25 and Fuzzy return in ~5ms; Semantic spends ~15ms; Hybrid (semantic + fuzzy) averages 52ms.
 - **Fuzzy baseline:** Recall@1 22%, highlighting the need for semantic signals on natural-language queries.
 
 ### With Reranker (`Alibaba-NLP/gte-reranker-modernbert-base`)
-![With Reranker](img/spider_benchmark_with_reranker.png)
+![With Reranker](https://raw.githubusercontent.com/Neehan/schema-search/refs/heads/main/img/spider_benchmark_with_reranker.png)
 - **Indexing:** 0.25s ± 0.05s per database (same 18 DBs).
 - **Accuracy:** All strategies converge around Recall@1 62% and MRR ≈ 0.92; Fuzzy jumps from 51% → 92% MRR.
 - **Latency trade-off:** Extra CrossEncoder pass lifts per-query latency to ~0.18–0.29s depending on strategy.
@@ -268,7 +267,7 @@ search = SchemaSearch(
 5. **Optional reranking** with CrossEncoder to refine results
 6. Return top tables with full schema and relationships
 
-Cache stored in `.schema_search_cache/` (configurable in `config.yml`)
+Cache stored in `/tmp/.schema_search_cache/` (configurable in `config.yml`)
 
 ## License
 

{schema_search-0.1.4 → schema_search-0.1.6}/README.md

@@ -159,14 +159,14 @@ We [benchmarked](/tests/test_spider_eval.py) on the Spider dataset (1,234 train
 **Memory:** The embedding model requires ~90 MB and the optional reranker adds ~155 MB. Actual process memory depends on your Python runtime.
 
 ### Without Reranker (`reranker.model: null`)
-![Without Reranker](img/spider_benchmark_without_reranker.png)
+![Without Reranker](https://raw.githubusercontent.com/Neehan/schema-search/refs/heads/main/img/spider_benchmark_without_reranker.png)
 - **Indexing:** 0.22s ± 0.08s per database (18 total).
 - **Accuracy:** Hybrid leads with Recall@1 62% / MRR 0.93; Semantic follows at Recall@1 58% / MRR 0.89.
 - **Latency:** BM25 and Fuzzy return in ~5ms; Semantic spends ~15ms; Hybrid (semantic + fuzzy) averages 52ms.
 - **Fuzzy baseline:** Recall@1 22%, highlighting the need for semantic signals on natural-language queries.
 
 ### With Reranker (`Alibaba-NLP/gte-reranker-modernbert-base`)
-![With Reranker](img/spider_benchmark_with_reranker.png)
+![With Reranker](https://raw.githubusercontent.com/Neehan/schema-search/refs/heads/main/img/spider_benchmark_with_reranker.png)
 - **Indexing:** 0.25s ± 0.05s per database (same 18 DBs).
 - **Accuracy:** All strategies converge around Recall@1 62% and MRR ≈ 0.92; Fuzzy jumps from 51% → 92% MRR.
 - **Latency trade-off:** Extra CrossEncoder pass lifts per-query latency to ~0.18–0.29s depending on strategy.
@@ -216,7 +216,7 @@ search = SchemaSearch(
 5. **Optional reranking** with CrossEncoder to refine results
 6. Return top tables with full schema and relationships
 
-Cache stored in `.schema_search_cache/` (configurable in `config.yml`)
+Cache stored in `/tmp/.schema_search_cache/` (configurable in `config.yml`)
 
 ## License
 

{schema_search-0.1.4 → schema_search-0.1.6/schema_search.egg-info}/PKG-INFO

@@ -1,18 +1,16 @@
 Metadata-Version: 2.4
 Name: schema-search
-Version: 0.1.4
-Summary: Natural language search for database schemas with graph-aware semantic retrieval
-Home-page: https://github.com/neehan/schema-search
-Author: 
+Version: 0.1.6
+Summary: Natural language database schema search with graph-aware semantic retrieval
+Home-page: https://adibhasan.com/blog/schema-search/
+Author: Adib Hasan
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.8
-Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
-Requires-Python: >=3.8
+Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: sqlalchemy>=1.4.0
@@ -40,6 +38,7 @@ Requires-Dist: snowflake-sqlalchemy>=1.4.0; extra == "snowflake"
 Requires-Dist: snowflake-connector-python>=3.0.0; extra == "snowflake"
 Provides-Extra: bigquery
 Requires-Dist: sqlalchemy-bigquery>=1.6.0; extra == "bigquery"
+Dynamic: author
 Dynamic: classifier
 Dynamic: description
 Dynamic: description-content-type
@@ -211,14 +210,14 @@ We [benchmarked](/tests/test_spider_eval.py) on the Spider dataset (1,234 train
 **Memory:** The embedding model requires ~90 MB and the optional reranker adds ~155 MB. Actual process memory depends on your Python runtime.
 
 ### Without Reranker (`reranker.model: null`)
-![Without Reranker](img/spider_benchmark_without_reranker.png)
+![Without Reranker](https://raw.githubusercontent.com/Neehan/schema-search/refs/heads/main/img/spider_benchmark_without_reranker.png)
 - **Indexing:** 0.22s ± 0.08s per database (18 total).
 - **Accuracy:** Hybrid leads with Recall@1 62% / MRR 0.93; Semantic follows at Recall@1 58% / MRR 0.89.
 - **Latency:** BM25 and Fuzzy return in ~5ms; Semantic spends ~15ms; Hybrid (semantic + fuzzy) averages 52ms.
 - **Fuzzy baseline:** Recall@1 22%, highlighting the need for semantic signals on natural-language queries.
 
 ### With Reranker (`Alibaba-NLP/gte-reranker-modernbert-base`)
-![With Reranker](img/spider_benchmark_with_reranker.png)
+![With Reranker](https://raw.githubusercontent.com/Neehan/schema-search/refs/heads/main/img/spider_benchmark_with_reranker.png)
 - **Indexing:** 0.25s ± 0.05s per database (same 18 DBs).
 - **Accuracy:** All strategies converge around Recall@1 62% and MRR ≈ 0.92; Fuzzy jumps from 51% → 92% MRR.
 - **Latency trade-off:** Extra CrossEncoder pass lifts per-query latency to ~0.18–0.29s depending on strategy.
@@ -268,7 +267,7 @@ search = SchemaSearch(
 5. **Optional reranking** with CrossEncoder to refine results
 6. Return top tables with full schema and relationships
 
-Cache stored in `.schema_search_cache/` (configurable in `config.yml`)
+Cache stored in `/tmp/.schema_search_cache/` (configurable in `config.yml`)
 
 ## License
 

{schema_search-0.1.4 → schema_search-0.1.6}/setup.py

@@ -2,12 +2,12 @@ from setuptools import setup, find_packages
 
 setup(
     name="schema-search",
-    version="0.1.4",
-    description="Natural language search for database schemas with graph-aware semantic retrieval",
-    author="",
+    version="0.1.6",
+    description="Natural language database schema search with graph-aware semantic retrieval",
+    author="Adib Hasan",
     long_description=open("README.md").read(),
     long_description_content_type="text/markdown",
-    url="https://github.com/neehan/schema-search",
+    url="https://adibhasan.com/blog/schema-search/",
     packages=find_packages(),
     install_requires=[
         "sqlalchemy>=1.4.0",
@@ -49,14 +49,12 @@ setup(
             "schema-search=schema_search.mcp_server:main",
         ],
     },
-    python_requires=">=3.8",
+    python_requires=">=3.10",
     classifiers=[
         "Development Status :: 3 - Alpha",
         "Intended Audience :: Developers",
         "License :: OSI Approved :: MIT License",
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.8",
-        "Programming Language :: Python :: 3.9",
         "Programming Language :: Python :: 3.10",
         "Programming Language :: Python :: 3.11",
     ],