From ab3d1a3e2d439ef0c558bd883736fca7d33059ce Mon Sep 17 00:00:00 2001
From: anakin87 <stefanofiorucci@gmail.com>
Date: Tue, 16 Apr 2024 17:49:26 +0100
Subject: [PATCH] Qdrant: add embedding retrieval example

---
 .../qdrant/examples/embedding_retrieval.py    | 52 +++++++++++++++++++
 integrations/qdrant/pyproject.toml            |  2 +
 2 files changed, 54 insertions(+)
 create mode 100644 integrations/qdrant/examples/embedding_retrieval.py

diff --git a/integrations/qdrant/examples/embedding_retrieval.py b/integrations/qdrant/examples/embedding_retrieval.py
new file mode 100644
index 000000000..f009191e7
--- /dev/null
+++ b/integrations/qdrant/examples/embedding_retrieval.py
@@ -0,0 +1,52 @@
+# Install required packages for this example, including qdrant-haystack and other libraries needed
+# for Markdown conversion and embeddings generation. Use the following command:
+# pip install qdrant-haystack markdown-it-py mdit_plain sentence-transformers
+
+# Download some Markdown files to index.
+# git clone https://github.com/anakin87/neural-search-pills
+
+import glob
+
+from haystack import Pipeline
+from haystack.components.converters import MarkdownToDocument
+from haystack.components.embedders import SentenceTransformersDocumentEmbedder, SentenceTransformersTextEmbedder
+from haystack.components.preprocessors import DocumentSplitter
+from haystack.components.writers import DocumentWriter
+from haystack_integrations.components.retrievers.qdrant import QdrantEmbeddingRetriever
+from haystack_integrations.document_stores.qdrant import QdrantDocumentStore
+
+# Initialize QdrantDocumentStore: for simplicity, we use an in-memory store here.
+# You can also run a Qdrant instance using Docker or use Qdrant Cloud.
+document_store = QdrantDocumentStore(
+    ":memory:",
+    index="Document",
+    embedding_dim=768,
+    recreate_index=True,
+)
+
+# Create the indexing Pipeline and index some documents
+file_paths = glob.glob("neural-search-pills/pills/*.md")
+
+
+indexing = Pipeline()
+indexing.add_component("converter", MarkdownToDocument())
+indexing.add_component("splitter", DocumentSplitter(split_by="sentence", split_length=2))
+indexing.add_component("embedder", SentenceTransformersDocumentEmbedder())
+indexing.add_component("writer", DocumentWriter(document_store))
+indexing.connect("converter", "splitter")
+indexing.connect("splitter", "embedder")
+indexing.connect("embedder", "writer")
+
+indexing.run({"converter": {"sources": file_paths}})
+
+# Create the querying Pipeline and try a query
+querying = Pipeline()
+querying.add_component("embedder", SentenceTransformersTextEmbedder())
+querying.add_component("retriever", QdrantEmbeddingRetriever(document_store=document_store, top_k=3))
+querying.connect("embedder", "retriever")
+
+results = querying.run({"embedder": {"text": "What is a cross-encoder?"}})
+
+for doc in results["retriever"]["documents"]:
+    print(doc)
+    print("-" * 10)
diff --git a/integrations/qdrant/pyproject.toml b/integrations/qdrant/pyproject.toml
index a566de955..fc1ea7b7e 100644
--- a/integrations/qdrant/pyproject.toml
+++ b/integrations/qdrant/pyproject.toml
@@ -127,6 +127,8 @@ ban-relative-imports = "parents"
 [tool.ruff.per-file-ignores]
 # Tests can use magic values, assertions, and relative imports
 "tests/**/*" = ["PLR2004", "S101", "TID252"]
+# examples can contain "print" commands
+"examples/**/*" = ["T201"]
 
 
 [tool.coverage.run]