visual-rag-toolkit 0.1.3__tar.gz → 0.1.4__tar.gz

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: visual-rag-toolkit
-Version: 0.1.3
+Version: 0.1.4
 Summary: End-to-end visual document retrieval with ColPali, featuring two-stage pooling for scalable search
 Project-URL: Homepage, https://github.com/Ara-Yeroyan/visual-rag-toolkit
 Project-URL: Documentation, https://github.com/Ara-Yeroyan/visual-rag-toolkit#readme
@@ -88,9 +88,12 @@ Description-Content-Type: text/markdown
 [![PyPI](https://img.shields.io/pypi/v/visual-rag-toolkit)](https://pypi.org/project/visual-rag-toolkit/)
 [![Python](https://img.shields.io/pypi/pyversions/visual-rag-toolkit)](https://pypi.org/project/visual-rag-toolkit/)
 [![License](https://img.shields.io/pypi/l/visual-rag-toolkit)](LICENSE)
+[![Demo](https://img.shields.io/badge/Demo-Hugging%20Face-yellow)](https://huggingface.co/spaces/Yeroyan/visual-rag-toolkit)
 
 End-to-end visual document retrieval toolkit featuring **fast multi-stage retrieval** (prefetch with pooled vectors + exact MaxSim reranking).
 
+**[Try the Live Demo](https://huggingface.co/spaces/Yeroyan/visual-rag-toolkit)** - Upload PDFs, index to Qdrant, and query with visual retrieval.
+
 This repo contains:
 - a **Python package** (`visual_rag`)
 - a **Streamlit demo app** (`demo/`)
@@ -368,7 +371,7 @@ If you use this toolkit in your research, please cite:
 
 ```bibtex
 @software{visual_rag_toolkit,
-  title = {Visual RAG Toolkit: Scalable Visual Document Retrieval with Two-Stage Pooling},
+  title = {Visual RAG Toolkit: Scalable Visual Document Retrieval with 1D Convolutional Pooling},
   author = {Ara Yeroyan},
   year = {2026},
   url = {https://github.com/Ara-Yeroyan/visual-rag-toolkit}

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/README.md

@@ -3,9 +3,12 @@
 [![PyPI](https://img.shields.io/pypi/v/visual-rag-toolkit)](https://pypi.org/project/visual-rag-toolkit/)
 [![Python](https://img.shields.io/pypi/pyversions/visual-rag-toolkit)](https://pypi.org/project/visual-rag-toolkit/)
 [![License](https://img.shields.io/pypi/l/visual-rag-toolkit)](LICENSE)
+[![Demo](https://img.shields.io/badge/Demo-Hugging%20Face-yellow)](https://huggingface.co/spaces/Yeroyan/visual-rag-toolkit)
 
 End-to-end visual document retrieval toolkit featuring **fast multi-stage retrieval** (prefetch with pooled vectors + exact MaxSim reranking).
 
+**[Try the Live Demo](https://huggingface.co/spaces/Yeroyan/visual-rag-toolkit)** - Upload PDFs, index to Qdrant, and query with visual retrieval.
+
 This repo contains:
 - a **Python package** (`visual_rag`)
 - a **Streamlit demo app** (`demo/`)
@@ -283,7 +286,7 @@ If you use this toolkit in your research, please cite:
 
 ```bibtex
 @software{visual_rag_toolkit,
-  title = {Visual RAG Toolkit: Scalable Visual Document Retrieval with Two-Stage Pooling},
+  title = {Visual RAG Toolkit: Scalable Visual Document Retrieval with 1D Convolutional Pooling},
   author = {Ara Yeroyan},
   year = {2026},
   url = {https://github.com/Ara-Yeroyan/visual-rag-toolkit}

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/demo/__init__.py

@@ -7,4 +7,4 @@ A Streamlit-based UI for:
 - Interactive playground for visual search
 """
 
-__version__ = "0.1.0"
+__version__ = "0.1.4"

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/demo/ui/sidebar.py

@@ -16,6 +16,28 @@ from demo.qdrant_utils import (
 
 
 def render_sidebar():
+    # CSS to make sidebar metrics smaller
+    st.markdown("""
+    <style>
+    /* Smaller metrics in sidebar */
+    [data-testid="stSidebar"] [data-testid="stMetricValue"] {
+        font-size: 1.2rem !important;
+    }
+    [data-testid="stSidebar"] [data-testid="stMetricLabel"] {
+        font-size: 0.75rem !important;
+    }
+    /* Smaller expander headers in sidebar */
+    [data-testid="stSidebar"] [data-testid="stExpander"] summary {
+        font-size: 0.9rem !important;
+    }
+    /* Compact subheaders */
+    [data-testid="stSidebar"] h3 {
+        font-size: 1rem !important;
+        margin-bottom: 0.5rem !important;
+    }
+    </style>
+    """, unsafe_allow_html=True)
+    
     with st.sidebar:
         st.subheader("🔑 Qdrant Credentials")
         

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/demo/ui/upload.py

@@ -449,7 +449,7 @@ def process_pdfs(uploaded_files, config):
         
         if total_uploaded > 0:
             st.session_state["upload_success"] = f"Uploaded {total_uploaded} pages to {collection_name}"
-            st.balloons()
+            st.rerun()  # Immediately refresh to show success toast + balloons
             
     except Exception as e:
         st.error(f"❌ Processing error: {e}")

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "visual-rag-toolkit"
-version = "0.1.3"
+version = "0.1.4"
 description = "End-to-end visual document retrieval with ColPali, featuring two-stage pooling for scalable search"
 readme = "README.md"
 license = {file = "LICENSE"}
@@ -151,3 +151,27 @@ python_version = "3.9"
 warn_return_any = true
 warn_unused_configs = true
 ignore_missing_imports = true
+
+[tool.bumpversion]
+current_version = "0.1.4"
+parse = "(?P<major>\\d+)\\.(?P<minor>\\d+)\\.(?P<patch>\\d+)"
+serialize = ["{major}.{minor}.{patch}"]
+commit = true
+tag = true
+tag_name = "v{new_version}"
+message = "Bump version: {current_version} → {new_version}"
+
+[[tool.bumpversion.files]]
+filename = "pyproject.toml"
+search = 'version = "{current_version}"'
+replace = 'version = "{new_version}"'
+
+[[tool.bumpversion.files]]
+filename = "visual_rag/__init__.py"
+search = '__version__ = "{current_version}"'
+replace = '__version__ = "{new_version}"'
+
+[[tool.bumpversion.files]]
+filename = "demo/__init__.py"
+search = '__version__ = "{current_version}"'
+replace = '__version__ = "{new_version}"'

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/visual_rag/__init__.py

@@ -33,17 +33,17 @@ Each component works independently - use only what you need.
 
 import logging
 
-__version__ = "0.1.3"
+__version__ = "0.1.4"
 
 
 def setup_logging(level: str = "INFO", format: str = None) -> None:
     """
     Configure logging for visual_rag package.
-    
+
     Args:
         level: Log level ("DEBUG", "INFO", "WARNING", "ERROR")
         format: Custom format string. Default shows time, level, and message.
-    
+
     Example:
         >>> import visual_rag
         >>> visual_rag.setup_logging("INFO")
@@ -51,13 +51,13 @@ def setup_logging(level: str = "INFO", format: str = None) -> None:
     """
     if format is None:
         format = "[%(asctime)s] %(levelname)s - %(message)s"
-    
+
     logging.basicConfig(
         level=getattr(logging, level.upper(), logging.INFO),
         format=format,
         datefmt="%H:%M:%S",
     )
-    
+
     # Also set the visual_rag logger specifically
     logger = logging.getLogger("visual_rag")
     logger.setLevel(getattr(logging, level.upper(), logging.INFO))
@@ -111,13 +111,16 @@ try:
 except ImportError:
     QdrantAdmin = None
 
-try:
-    from visual_rag.demo_runner import demo
-except ImportError:
-    demo = None
+# demo is lazily imported to avoid RuntimeWarning when running as __main__
+# Access via visual_rag.demo() which triggers __getattr__
 
 # Config utilities (always available)
-from visual_rag.config import get, get_section, load_config
+try:
+    from visual_rag.config import get, get_section, load_config
+except ImportError:
+    get = None
+    get_section = None
+    load_config = None
 
 __all__ = [
     # Version
@@ -138,3 +141,15 @@ __all__ = [
     # Logging
     "setup_logging",
 ]
+
+
+def __getattr__(name: str):
+    """Lazy import for demo to avoid RuntimeWarning when running as __main__."""
+    if name == "demo":
+        try:
+            from visual_rag.demo_runner import demo
+
+            return demo
+        except ImportError:
+            return None
+    raise AttributeError(f"module 'visual_rag' has no attribute {name!r}")

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/visual_rag/demo_runner.py

@@ -52,13 +52,11 @@ def demo(
     cmd = [sys.executable, "-m", "streamlit", "run", str(app_path)]
     cmd += ["--server.address", str(host)]
     cmd += ["--server.port", str(int(port))]
-    cmd += ["--server.headless", "true" if headless else "false"]
+    # headless=true prevents browser from auto-opening; open_browser overrides
+    should_be_headless = headless and not open_browser
+    cmd += ["--server.headless", "true" if should_be_headless else "false"]
     cmd += ["--browser.gatherUsageStats", "false"]
     cmd += ["--server.runOnSave", "false"]
-    cmd += ["--browser.serverAddress", str(host)]
-    if not open_browser:
-        cmd += ["--browser.serverPort", str(int(port))]
-        cmd += ["--browser.open", "false"]
 
     if extra_args:
         cmd += list(extra_args)

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/visual_rag/indexing/qdrant_indexer.py

@@ -335,7 +335,9 @@ class QdrantIndexer:
                 return val.tolist()
             return val
 
-        def _build_qdrant_points(batch_points: List[Dict[str, Any]]) -> List[qdrant_models.PointStruct]:
+        def _build_qdrant_points(
+            batch_points: List[Dict[str, Any]],
+        ) -> List[qdrant_models.PointStruct]:
             qdrant_points: List[qdrant_models.PointStruct] = []
             for p in batch_points:
                 global_pooled = p.get("global_pooled_embedding")
@@ -382,7 +384,7 @@ class QdrantIndexer:
                 )
 
                 logger.info(f"   ✅ Uploaded {len(points)} points to Qdrant")
-                
+
                 if delay_between_batches > 0:
                     if _is_cancelled():
                         return 0

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/visual_rag/retrieval/multi_vector.py

@@ -56,14 +56,10 @@ class MultiVectorRetriever:
                 )
 
             qdrant_url = (
-                qdrant_url
-                or os.getenv("QDRANT_URL")
-                or os.getenv("SIGIR_QDRANT_URL")  # legacy
+                qdrant_url or os.getenv("QDRANT_URL") or os.getenv("SIGIR_QDRANT_URL")  # legacy
             )
             if not qdrant_url:
-                raise ValueError(
-                    "QDRANT_URL is required (pass qdrant_url or set env var)."
-                )
+                raise ValueError("QDRANT_URL is required (pass qdrant_url or set env var).")
 
             qdrant_api_key = (
                 qdrant_api_key
@@ -97,7 +93,10 @@ class MultiVectorRetriever:
                     _ = client.get_collections()
                 except Exception as e:
                     msg = str(e)
-                    if "StatusCode.PERMISSION_DENIED" in msg or "http2 header with status: 403" in msg:
+                    if (
+                        "StatusCode.PERMISSION_DENIED" in msg
+                        or "http2 header with status: 403" in msg
+                    ):
                         client = _make_client(False)
                     else:
                         raise

{visual_rag_toolkit-0.1.3 → visual_rag_toolkit-0.1.4}/visual_rag/retrieval/two_stage.py

@@ -22,7 +22,6 @@ from typing import Any, Dict, List, Optional, Union
 
 import numpy as np
 import torch
-
 from qdrant_client.http import models as qdrant_models
 from qdrant_client.models import FieldCondition, Filter, MatchAny, MatchValue
 

visual_rag_toolkit-0.1.3/demo/example_metadata_mapping_sigir.json

@@ -1,37 +0,0 @@
-{
-  "filenames": {
-    "sigir2025-llms": {
-      "year": 2025,
-      "source": "Conference Paper",
-      "district": null,
-      "doc_type": "paper",
-      "project": "sigir-demo",
-      "tags": ["llms", "retrieval"]
-    },
-    "sigir2025-ginger": {
-      "year": 2025,
-      "source": "Conference Paper",
-      "district": null,
-      "doc_type": "paper",
-      "project": "sigir-demo",
-      "tags": ["ginger", "case-study"]
-    },
-    "2505.15859v1": {
-      "year": 2025,
-      "source": "arXiv",
-      "district": null,
-      "doc_type": "preprint",
-      "project": "sigir-demo",
-      "tags": ["arxiv", "ranking"]
-    },
-    "2507.04942v2": {
-      "year": 2025,
-      "source": "arXiv",
-      "district": null,
-      "doc_type": "preprint",
-      "project": "sigir-demo",
-      "tags": ["arxiv", "rag"]
-    }
-  }
-}
-