From 9c218af13a9f641d33af6531d22e1e300e859929 Mon Sep 17 00:00:00 2001 From: davidmyriel Date: Tue, 26 Nov 2024 14:58:31 -0800 Subject: [PATCH] fix semantic --- .../search-precision/reranking-hybrid-search.md | 2 +- ...ector-search.md => reranking-semantic-search.md} | 10 +++++----- .../image1.png | Bin .../image5.png | Bin .../image6.png | Bin .../image7.png | Bin 6 files changed, 6 insertions(+), 6 deletions(-) rename qdrant-landing/content/documentation/search-precision/{reranking-vector-search.md => reranking-semantic-search.md} (98%) rename qdrant-landing/static/documentation/examples/{reranking-vector-search => reranking-semantic-search}/image1.png (100%) rename qdrant-landing/static/documentation/examples/{reranking-vector-search => reranking-semantic-search}/image5.png (100%) rename qdrant-landing/static/documentation/examples/{reranking-vector-search => reranking-semantic-search}/image6.png (100%) rename qdrant-landing/static/documentation/examples/{reranking-vector-search => reranking-semantic-search}/image7.png (100%) diff --git a/qdrant-landing/content/documentation/search-precision/reranking-hybrid-search.md b/qdrant-landing/content/documentation/search-precision/reranking-hybrid-search.md index 149a42a97..16722ad13 100644 --- a/qdrant-landing/content/documentation/search-precision/reranking-hybrid-search.md +++ b/qdrant-landing/content/documentation/search-precision/reranking-hybrid-search.md @@ -4,7 +4,7 @@ weight: 2 partition: build --- -# Implementing Hybrid Search with Reranking +# Reranking Hybrid Search Results with Qdrant Vector Database Hybrid search combines dense and sparse retrieval to deliver precise and comprehensive results. By adding reranking with ColBERT, you can further refine search outputs for maximum relevance. diff --git a/qdrant-landing/content/documentation/search-precision/reranking-vector-search.md b/qdrant-landing/content/documentation/search-precision/reranking-semantic-search.md similarity index 98% rename from qdrant-landing/content/documentation/search-precision/reranking-vector-search.md rename to qdrant-landing/content/documentation/search-precision/reranking-semantic-search.md index dcd2d52b8..f22bfabe5 100644 --- a/qdrant-landing/content/documentation/search-precision/reranking-vector-search.md +++ b/qdrant-landing/content/documentation/search-precision/reranking-semantic-search.md @@ -1,5 +1,5 @@ --- -title: Reranking in Vector Search +title: Reranking in Semantic Search weight: 1 partition: build --- @@ -20,7 +20,7 @@ This section is broken down into key parts to help you easily grasp the backgrou In search systems, two metrics—precision and recall—are the backbone of success. But what do they mean? Precision tells us how many of the retrieved results are actually relevant, while recall measures how well we’ve captured all the relevant results out there. Simply put: -![image5.png](/documentation/examples/reranking-vector-search/image5.png) +![image5.png](/documentation/examples/reranking-semantic-search/image5.png) Sparse vector searches usually give you high precision because they’re great at finding exact matches. But, here's the catch—your recall can suffer when relevant documents don’t contain those exact keywords. On the flip side, dense vector searches are fantastic for recall since they grasp the broader, semantic meaning of your query. However, this can lead to lower precision, where you might see results that are only loosely related. @@ -32,7 +32,7 @@ Picture this: You walk into a massive library and ask for a book on "climate cha Now, imagine a smarter, more intuitive librarian who really gets what you’re after. This one knows exactly which books are most impactful, the most current, and perfectly aligned with what you need. That’s what reranking does for your search results—it doesn’t just grab any relevant document; it smartly reorders them so the best ones land at the top of your list. It’s like having a librarian who knows exactly what you’re looking for before you do! -![image6.png](/documentation/examples/reranking-vector-search/image6.png) +![image6.png](/documentation/examples/reranking-semantic-search/image6.png) An illustration of the rerank model prioritizing better results @@ -68,7 +68,7 @@ A typical search system works in two main stages: Ingestion and Retrieval. Think Check out the architectural diagram below to visualize how these stages work together. -![image1.png](/documentation/examples/reranking-vector-search/image1.png) +![image1.png](/documentation/examples/reranking-semantic-search/image1.png) The two essential stages of a search system: Ingestion and Retrieval Process @@ -125,7 +125,7 @@ Qdrant is a powerful vector similarity search engine that gives you a production Finally, under the Overview section, you’ll see the following code snippet: -![image7.png](/documentation/examples/reranking-vector-search/image7.png) +![image7.png](/documentation/examples/reranking-semantic-search/image7.png) Qdrant Overview Section diff --git a/qdrant-landing/static/documentation/examples/reranking-vector-search/image1.png b/qdrant-landing/static/documentation/examples/reranking-semantic-search/image1.png similarity index 100% rename from qdrant-landing/static/documentation/examples/reranking-vector-search/image1.png rename to qdrant-landing/static/documentation/examples/reranking-semantic-search/image1.png diff --git a/qdrant-landing/static/documentation/examples/reranking-vector-search/image5.png b/qdrant-landing/static/documentation/examples/reranking-semantic-search/image5.png similarity index 100% rename from qdrant-landing/static/documentation/examples/reranking-vector-search/image5.png rename to qdrant-landing/static/documentation/examples/reranking-semantic-search/image5.png diff --git a/qdrant-landing/static/documentation/examples/reranking-vector-search/image6.png b/qdrant-landing/static/documentation/examples/reranking-semantic-search/image6.png similarity index 100% rename from qdrant-landing/static/documentation/examples/reranking-vector-search/image6.png rename to qdrant-landing/static/documentation/examples/reranking-semantic-search/image6.png diff --git a/qdrant-landing/static/documentation/examples/reranking-vector-search/image7.png b/qdrant-landing/static/documentation/examples/reranking-semantic-search/image7.png similarity index 100% rename from qdrant-landing/static/documentation/examples/reranking-vector-search/image7.png rename to qdrant-landing/static/documentation/examples/reranking-semantic-search/image7.png