rust/lancedb
1
0
Fork 0
mirror of https://github.com/lancedb/lancedb.git synced 2026-10-12 14:52:26 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
07e8a8db17155850469a06a464a3883efb480f90
lancedb/docs/web/snippets/integrations.mdx
T
Brendan Clement 07e8a8db17 docs: fix copyable Python examples that fail as published 
- Full-text search: the multiline query chain was not parenthesized, so
  pasting it raised an IndentationError.
- Pandas async: the example chained .select onto the coroutine an async
  search returns; it now awaits the search first, and its test asserts
  the result.
- Pydantic: the example searched with a string, which needs a full-text
  index or an embedding function the table does not have; it searches
  with a vector, as the page describes, and asserts the model returned.
- PyTorch: the column-selection example selected columns the page's
  table lacks and printed a schema from a batch that is a dict. It now
  builds its own table and prints the batch's columns.
- Data evolution: the setup used the test's tmp_db fixture, an optional
  data argument and an unimported numpy. It now needs only `db`, which
  the page says how to open.
2026-10-09 13:36:49 -07:00

183 lines
50 KiB
Plaintext
Raw Blame History

{/* Auto-generated by scripts/mdx_snippets_gen.py. Do not edit manually. */}
export const PyEmbeddingAwsUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\nmodel = get_registry().get(\"bedrock-text\").create()\n\nclass TextModel(LanceModel):\n text: str = model.SourceField()\n vector: Vector(model.ndims()) = model.VectorField()\n\ndf = pd.DataFrame({\"text\": [\"hello world\", \"goodbye world\"]})\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"bedrock-demo\"))\ntbl = db.create_table(\"test\", schema=TextModel, mode=\"overwrite\")\n\ntbl.add(df)\nrs = tbl.search(\"hello\").limit(1).to_pandas()\nprint(rs.head())\n";
export const PyEmbeddingCohereUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import EmbeddingFunctionRegistry\nfrom lancedb.pydantic import LanceModel, Vector\n\ncohere = (\n EmbeddingFunctionRegistry.get_instance()\n .get(\"cohere\")\n .create(name=\"embed-multilingual-v2.0\")\n)\n\nclass TextModel(LanceModel):\n text: str = cohere.SourceField()\n vector: Vector(cohere.ndims()) = cohere.VectorField()\n\ndata = [{\"text\": \"hello world\"}, {\"text\": \"goodbye world\"}]\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"cohere-demo\"))\ntbl = db.create_table(\"test\", schema=TextModel, mode=\"overwrite\")\ntbl.add(data)\n";
export const PyEmbeddingColpaliSetup = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\nimport requests\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, MultiVector\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"colpali-demo\"))\nfunc = get_registry().get(\"colpali\").create()\n\nclass Images(LanceModel):\n label: str\n image_uri: str = func.SourceField()\n image_bytes: bytes = func.SourceField()\n vector: MultiVector(func.ndims()) = func.VectorField()\n vec_from_bytes: MultiVector(func.ndims()) = func.VectorField()\n\ntable = db.create_table(\"images\", schema=Images)\nlabels = [\"cat\", \"dog\", \"horse\"]\nuris = [\n \"http://farm1.staticflickr.com/53/167798175_7c7845bbbd_z.jpg\",\n \"http://farm9.staticflickr.com/8387/8602747737_2e5c2a45d4_z.jpg\",\n \"http://farm9.staticflickr.com/8216/8434969557_d37882c42d_z.jpg\",\n]\nimage_bytes = [requests.get(uri).content for uri in uris]\ntable.add(\n pd.DataFrame({\"label\": labels, \"image_uri\": uris, \"image_bytes\": image_bytes})\n)\n";
export const PyEmbeddingColpaliTextSearch = "actual = (\n table.search(\"a furry pet\", vector_column_name=\"vector\")\n .limit(1)\n .to_pydantic(Images)[0]\n)\nprint(actual.label)\n";
export const PyEmbeddingGeminiUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\nmodel = get_registry().get(\"gemini-text\").create()\n\nclass TextModel(LanceModel):\n text: str = model.SourceField()\n vector: Vector(model.ndims()) = model.VectorField()\n\ndf = pd.DataFrame({\"text\": [\"hello world\", \"goodbye world\"]})\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"gemini-demo\"))\ntbl = db.create_table(\"test\", schema=TextModel, mode=\"overwrite\")\n\ntbl.add(df)\nrs = tbl.search(\"hello\").limit(1).to_pandas()\nprint(rs.head())\n";
export const PyEmbeddingHuggingfaceUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"huggingface-demo\"))\nmodel = get_registry().get(\"huggingface\").create(name=\"facebook/bart-base\")\n\nclass Words(LanceModel):\n text: str = model.SourceField()\n vector: Vector(model.ndims()) = model.VectorField()\n\ndf = pd.DataFrame({\"text\": [\"hi hello sayonara\", \"goodbye world\"]})\ntable = db.create_table(\"greets\", schema=Words)\ntable.add(df)\nquery = \"old greeting\"\nactual = table.search(query).limit(1).to_pydantic(Words)[0]\nprint(actual.text)\n";
export const PyEmbeddingIbmUsage = "import os\nimport tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import EmbeddingFunctionRegistry\nfrom lancedb.pydantic import LanceModel, Vector\n\nwatsonx_embed = (\n EmbeddingFunctionRegistry.get_instance()\n .get(\"watsonx\")\n .create(\n name=\"ibm/slate-125m-english-rtrvr\",\n api_key=os.environ.get(\"WATSONX_API_KEY\"),\n project_id=os.environ.get(\"WATSONX_PROJECT_ID\"),\n )\n)\n\nclass TextModel(LanceModel):\n text: str = watsonx_embed.SourceField()\n vector: Vector(watsonx_embed.ndims()) = watsonx_embed.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"watsonx-demo\"))\ntbl = db.create_table(\"watsonx_test\", schema=TextModel, mode=\"overwrite\")\ntbl.add(data)\n\nrs = tbl.search(\"hello\").limit(1).to_pandas()\nprint(rs.head())\n";
export const PyEmbeddingImagebindAudioSearch = "query_audio = \"./assets/car_audio2.wav\"\nactual = table.search(query_audio).limit(1).to_pydantic(ImageBindModel)[0]\nprint(actual.text == \"car\")\n";
export const PyEmbeddingImagebindImageSearch = "query_image = \"./assets/dog_image2.jpg\"\nactual = table.search(query_image).limit(1).to_pydantic(ImageBindModel)[0]\nprint(actual.text == \"dog\")\n";
export const PyEmbeddingImagebindSetup = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ndb = lancedb.connect(\"/tmp/imagebind-db\")\nfunc = get_registry().get(\"imagebind\").create()\n\nclass ImageBindModel(LanceModel):\n text: str\n image_uri: str = func.SourceField()\n audio_path: str\n vector: Vector(func.ndims()) = func.VectorField()\n\ntext_list = [\"A dog.\", \"A car\", \"A bird\"]\nimage_paths = [\n \"./assets/dog_image.jpg\",\n \"./assets/car_image.jpg\",\n \"./assets/bird_image.jpg\",\n]\naudio_paths = [\n \"./assets/dog_audio.wav\",\n \"./assets/car_audio.wav\",\n \"./assets/bird_audio.wav\",\n]\n\ninputs = [\n {\"text\": a, \"audio_path\": b, \"image_uri\": c}\n for a, b, c in zip(text_list, audio_paths, image_paths)\n]\n\ntable = db.create_table(\"img_bind\", schema=ImageBindModel)\ntable.add(inputs)\n";
export const PyEmbeddingImagebindTextSearch = "query = \"an animal which flies and tweets\"\nactual = table.search(query).limit(1).to_pydantic(ImageBindModel)[0]\nprint(actual.text == \"bird\")\n";
export const PyEmbeddingInstructorUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ninstructor = (\n get_registry()\n .get(\"instructor\")\n .create(\n source_instruction=\"represent the document for retrieval\",\n query_instruction=\"represent the document for retrieving the most similar documents\",\n )\n)\n\nclass Schema(LanceModel):\n vector: Vector(instructor.ndims()) = instructor.VectorField()\n text: str = instructor.SourceField()\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"instructor-demo\"))\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\n\ntexts = [\n {\n \"text\": \"Capitalism has been dominant in the Western world since the end of feudalism.\"\n },\n {\n \"text\": \"The disparate impact theory is especially controversial under the Fair Housing Act.\"\n },\n {\n \"text\": \"Disparate impact in United States labor law refers to practices in employment.\"\n },\n]\n\ntbl.add(texts)\n";
export const PyEmbeddingJinaMultimodal = "import os\nimport tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\nimport requests\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\nos.environ[\"JINA_API_KEY\"] = os.environ.get(\"JINA_API_KEY\", \"jina_*\")\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"jina-images\"))\nfunc = get_registry().get(\"jina\").create()\n\nclass Images(LanceModel):\n label: str\n image_uri: str = func.SourceField()\n image_bytes: bytes = func.SourceField()\n vector: Vector(func.ndims()) = func.VectorField()\n vec_from_bytes: Vector(func.ndims()) = func.VectorField()\n\ntable = db.create_table(\"images\", schema=Images)\nlabels = [\"cat\", \"cat\", \"dog\", \"dog\", \"horse\", \"horse\"]\nuris = [\n \"http://farm1.staticflickr.com/53/167798175_7c7845bbbd_z.jpg\",\n \"http://farm1.staticflickr.com/134/332220238_da527d8140_z.jpg\",\n \"http://farm9.staticflickr.com/8387/8602747737_2e5c2a45d4_z.jpg\",\n \"http://farm5.staticflickr.com/4092/5017326486_1f46057f5f_z.jpg\",\n \"http://farm9.staticflickr.com/8216/8434969557_d37882c42d_z.jpg\",\n \"http://farm6.staticflickr.com/5142/5835678453_4f3a4edb45_z.jpg\",\n]\nimage_bytes = [requests.get(uri).content for uri in uris]\ntable.add(\n pd.DataFrame({\"label\": labels, \"image_uri\": uris, \"image_bytes\": image_bytes})\n)\n";
export const PyEmbeddingJinaText = "import os\nimport tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import EmbeddingFunctionRegistry\nfrom lancedb.pydantic import LanceModel, Vector\n\nos.environ[\"JINA_API_KEY\"] = os.environ[\"JINA_API_KEY\"]\n\njina_embed = (\n EmbeddingFunctionRegistry.get_instance()\n .get(\"jina\")\n .create(name=\"jina-embeddings-v2-base-en\")\n)\n\nclass TextModel(LanceModel):\n text: str = jina_embed.SourceField()\n vector: Vector(jina_embed.ndims()) = jina_embed.VectorField()\n\ndata = [{\"text\": \"hello world\"}, {\"text\": \"goodbye world\"}]\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"jina-text\"))\ntbl = db.create_table(\"test\", schema=TextModel, mode=\"overwrite\")\n\ntbl.add(data)\n";
export const PyEmbeddingOllamaUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"ollama-demo\"))\nfunc = get_registry().get(\"ollama\").create(name=\"nomic-embed-text\")\n\nclass Words(LanceModel):\n text: str = func.SourceField()\n vector: Vector(func.ndims()) = func.VectorField()\n\ntable = db.create_table(\"words\", schema=Words, mode=\"overwrite\")\ntable.add(\n [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n ]\n)\n\nquery = \"greetings\"\nactual = table.search(query).limit(1).to_pydantic(Words)[0]\nprint(actual.text)\n";
export const PyEmbeddingOpenaiBasic = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ndb_path = Path(tempfile.mkdtemp()) / \"openai-embeddings\"\ndb = lancedb.connect(str(db_path))\nfunc = get_registry().get(\"openai\").create(name=\"text-embedding-ada-002\")\n\nclass Words(LanceModel):\n text: str = func.SourceField()\n vector: Vector(func.ndims()) = func.VectorField()\n\ntable = db.create_table(\"words\", schema=Words, mode=\"overwrite\")\ntable.add(\n [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n ]\n)\n\nquery = \"greetings\"\nactual = table.search(query).limit(1).to_pydantic(Words)[0]\nprint(actual.text)\n";
export const PyEmbeddingOpenclipImageSearch = "import io\n\nfrom PIL import Image\n\nquery_image_uri = \"http://farm1.staticflickr.com/200/467715466_ed4a31801f_z.jpg\"\nimage_bytes = requests.get(query_image_uri).content\nquery_image = Image.open(io.BytesIO(image_bytes))\nactual = table.search(query_image).limit(1).to_pydantic(Images)[0]\nprint(actual.label == \"dog\")\n\nother = (\n table.search(query_image, vector_column_name=\"vec_from_bytes\")\n .limit(1)\n .to_pydantic(Images)[0]\n)\nprint(other.label)\n";
export const PyEmbeddingOpenclipSetup = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\nimport requests\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"openclip-demo\"))\nfunc = get_registry().get(\"open-clip\").create()\n\nclass Images(LanceModel):\n label: str\n image_uri: str = func.SourceField()\n image_bytes: bytes = func.SourceField()\n vector: Vector(func.ndims()) = func.VectorField()\n vec_from_bytes: Vector(func.ndims()) = func.VectorField()\n\ntable = db.create_table(\"images\", schema=Images)\nlabels = [\"cat\", \"cat\", \"dog\", \"dog\", \"horse\", \"horse\"]\nuris = [\n \"http://farm1.staticflickr.com/53/167798175_7c7845bbbd_z.jpg\",\n \"http://farm1.staticflickr.com/134/332220238_da527d8140_z.jpg\",\n \"http://farm9.staticflickr.com/8387/8602747737_2e5c2a45d4_z.jpg\",\n \"http://farm5.staticflickr.com/4092/5017326486_1f46057f5f_z.jpg\",\n \"http://farm9.staticflickr.com/8216/8434969557_d37882c42d_z.jpg\",\n \"http://farm6.staticflickr.com/5142/5835678453_4f3a4edb45_z.jpg\",\n]\nimage_bytes = [requests.get(uri).content for uri in uris]\ntable.add(\n pd.DataFrame({\"label\": labels, \"image_uri\": uris, \"image_bytes\": image_bytes})\n)\n";
export const PyEmbeddingOpenclipTextSearch = "actual = table.search(\"man's best friend\").limit(1).to_pydantic(Images)[0]\nprint(actual.label)\n\nfrombytes = (\n table.search(\"man's best friend\", vector_column_name=\"vec_from_bytes\")\n .limit(1)\n .to_pydantic(Images)[0]\n)\nprint(frombytes.label)\n";
export const PyEmbeddingSentenceTransformersBaai = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"sentence-transformers\"))\nmodel = (\n get_registry()\n .get(\"sentence-transformers\")\n .create(name=\"BAAI/bge-small-en-v1.5\", device=\"cpu\")\n)\n\nclass Words(LanceModel):\n text: str = model.SourceField()\n vector: Vector(model.ndims()) = model.VectorField()\n\ntable = db.create_table(\"words\", schema=Words)\ntable.add(\n [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n ]\n)\n\nquery = \"greetings\"\nactual = table.search(query).limit(1).to_pydantic(Words)[0]\nprint(actual.text)\n";
export const PyEmbeddingVoyageaiMultimodal = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import EmbeddingFunctionRegistry\nfrom lancedb.pydantic import LanceModel, Vector\n\n# Create multimodal embedding function with custom dimension\nvoyageai = (\n EmbeddingFunctionRegistry.get_instance()\n .get(\"voyageai\")\n .create(name=\"voyage-multimodal-3.5\", output_dimension=512)\n)\n\nclass ImageModel(LanceModel):\n image_uri: str = voyageai.SourceField()\n vector: Vector(voyageai.ndims()) = voyageai.VectorField()\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"voyageai-multimodal\"))\ntbl = db.create_table(\"images\", schema=ImageModel, mode=\"overwrite\")\n\n# Add images using URLs\ntbl.add(\n [\n {\"image_uri\": \"https://upload.wikimedia.org/wikipedia/commons/thumb/4/47/PNG_transparency_demonstration_1.png/300px-PNG_transparency_demonstration_1.png\"},\n ]\n)\n\n# Search with text query\nresults = tbl.search(\"dice\").limit(1).to_list()\nprint(results)\n";
export const PyEmbeddingVoyageaiUsage = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.embeddings import EmbeddingFunctionRegistry\nfrom lancedb.pydantic import LanceModel, Vector\n\nvoyageai = (\n EmbeddingFunctionRegistry.get_instance().get(\"voyageai\").create(name=\"voyage-3\")\n)\n\nclass TextModel(LanceModel):\n text: str = voyageai.SourceField()\n vector: Vector(voyageai.ndims()) = voyageai.VectorField()\n\ndata = [{\"text\": \"hello world\"}, {\"text\": \"goodbye world\"}]\n\ndb = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"voyageai-demo\"))\ntbl = db.create_table(\"test\", schema=TextModel, mode=\"overwrite\")\n\ntbl.add(data)\n";
export const PyFrameworksAgnoAgent = "agent = Agent(\n model=OpenAIResponses(id=\"gpt-5-mini\"),\n knowledge=knowledge,\n search_knowledge=True,\n instructions=\"Search the transcript and answer only from retrieved context.\",\n markdown=True,\n)\n";
export const PyFrameworksAgnoCliChat = "agent.print_response(\n \"Summarize the loaded video transcript in 5 concise bullet points.\",\n stream=True,\n)\nwhile True:\n question = input(\"You: \").strip()\n if question.lower() in {\"exit\", \"quit\", \"bye\"}:\n break\n agent.print_response(question, stream=True)\n";
export const PyFrameworksAgnoIngestYoutube = "youtube_url = \"https://www.youtube.com/watch?v=wl6mFyXoxos\"\nvideo_id = extract_video_id(youtube_url)\nytt = YouTubeTranscriptApi()\ntranscript_segments = ytt.fetch(video_id, languages=[\"en\", \"en-US\"]).to_raw_data()\ntranscript_text = \" \".join(segment[\"text\"] for segment in transcript_segments)\n\nknowledge.insert(\n name=f\"YouTube Transcript ({video_id})\",\n text_content=transcript_text,\n metadata={\"source\": \"youtube\", \"video_id\": video_id, \"video_url\": youtube_url},\n)\n";
export const PyFrameworksAgnoSetup = "import os\nimport re\n\nfrom agno.agent import Agent\nfrom agno.knowledge.embedder.openai import OpenAIEmbedder\nfrom agno.knowledge.knowledge import Knowledge\nfrom agno.models.openai import OpenAIResponses\nfrom agno.vectordb.lancedb import LanceDb, SearchType\nfrom youtube_transcript_api import YouTubeTranscriptApi\n\nif \"OPENAI_API_KEY\" not in os.environ:\n os.environ[\"OPENAI_API_KEY\"] = \"sk-...\"\n\ndef extract_video_id(youtube_url: str) -> str:\n match = re.search(r\"(?<=v=)[\\w-]+\", youtube_url) or re.search(\n r\"(?<=be/)[\\w-]+\", youtube_url\n )\n if not match:\n raise ValueError(\"Could not parse YouTube video ID from URL\")\n return match.group(0)\n\nknowledge = Knowledge(\n vector_db=LanceDb(\n uri=\"./tmp/lancedb\",\n table_name=\"youtube_transcripts\",\n search_type=SearchType.hybrid,\n embedder=OpenAIEmbedder(id=\"text-embedding-3-small\"),\n ),\n)\n";
export const PyFrameworksLangchainAddImages = "image_uris = [\"./assets/image-1.png\", \"./assets/image-2.png\"]\nvector_store.add_images(uris=image_uris)\n# here image_uris are local fs paths to the images.\n";
export const PyFrameworksLangchainAddTexts = "vector_store.add_texts(texts=[\"test_123\"], metadatas=[{\"source\": \"wiki\"}])\n\n# Additionally, to explore the table you can load it into a df or save it in a csv file:\n\ntbl = vector_store.get_table()\nprint(\"tbl:\", tbl)\npd_df = tbl.to_pandas()\npd_df.to_csv(\"docsearch.csv\", index=False)\n\n# you can also create a new vector store object using an older connection object:\nvector_store = LanceDB(connection=tbl, embedding=embeddings)\n";
export const PyFrameworksLangchainCreateIndex = "# for creating vector index\nvector_store.create_index(vector_col=\"vector\", metric=\"cosine\")\n\n# for creating scalar index(for non-vector columns)\nvector_store.create_index(col_name=\"text\")\n";
export const PyFrameworksLangchainMaxMarginalRelevance = "result = docsearch.max_marginal_relevance_search(query=\"text\")\nresult_texts = [doc.page_content for doc in result]\nprint(result_texts)\n\n# search by vector :\nresult = docsearch.max_marginal_relevance_search_by_vector(\n embeddings.embed_query(\"text\")\n)\nresult_texts = [doc.page_content for doc in result]\nprint(result_texts)\n";
export const PyFrameworksLangchainQuickStart = "import os\n\nfrom langchain.document_loaders import TextLoader\nfrom langchain.vectorstores import LanceDB\nfrom langchain_openai import OpenAIEmbeddings\nfrom langchain_text_splitters import CharacterTextSplitter\n\nos.environ[\"OPENAI_API_KEY\"] = \"sk-...\"\n\nloader = TextLoader(\n \"../../modules/state_of_the_union.txt\"\n) # Replace with your data path\ndocuments = loader.load()\n\ndocuments = CharacterTextSplitter().split_documents(documents)\nembeddings = OpenAIEmbeddings()\n\ndocsearch = LanceDB.from_documents(documents, embeddings)\nquery = \"What did the president say about Ketanji Brown Jackson\"\ndocs = docsearch.similarity_search(query)\nprint(docs[0].page_content)\n";
export const PyFrameworksLangchainSimilaritySearch = "docs = docsearch.similarity_search(query)\nprint(docs[0].page_content)\n";
export const PyFrameworksLangchainSimilaritySearchByVector = "docs = docsearch.similarity_search_by_vector(query)\nprint(docs[0].page_content)\n";
export const PyFrameworksLangchainSimilaritySearchByVectorWithScores = "query_embedding = embeddings.embed_query(\"text\")\ndocs = docsearch.similarity_search_by_vector_with_relevance_scores(query_embedding)\nprint(\"relevance score - \", docs[0][1])\nprint(\"text- \", docs[0][0].page_content[:1000])\n";
export const PyFrameworksLangchainSimilaritySearchWithScores = "docs = docsearch.similarity_search_with_relevance_scores(query)\nprint(\"relevance score - \", docs[0][1])\nprint(\"text- \", docs[0][0].page_content[:1000])\n";
export const PyFrameworksLangchainVectorStoreConfig = "db_url = \"db://lang_test\" # url of db you created\napi_key = \"xxxxx\" # your API key\nregion = \"us-east-1-dev\" # your selected region\n\nvector_store = LanceDB(\n uri=db_url,\n api_key=api_key, # (dont include for local API)\n region=region, # (dont include for local API)\n embedding=embeddings,\n table_name=\"langchain_test\", # Optional\n)\n";
export const PyFrameworksLerobotFilterFrames = "frame_rows = (\n frames.search()\n .where(\"episode_index = 0 AND frame_index < 10\", prefilter=True)\n .select([\"episode_index\", \"frame_index\", \"timestamp\", \"action\"])\n .limit(10)\n .to_list()\n)\n\nfor row in frame_rows:\n print(row[\"episode_index\"], row[\"frame_index\"], row[\"timestamp\"])\n";
export const PyFrameworksLerobotLoad = "from lerobot.datasets import LeRobotDataset\n\n# Hub dataset repo: storage_format in meta/info.json routes to the Lance reader\ndataset = LeRobotDataset(\"lance-format/pusht-lance\")\n\n# HF Storage Bucket, read in place, nothing downloaded\ndataset = LeRobotDataset(\"lance-format/lerobot-tests/pusht-lance\", repo_type=\"bucket\")\n\nsample = dataset[100]\nprint(sample[\"observation.state\"].shape)\nprint(sample[\"observation.image\"].shape)\n";
export const PyFrameworksLerobotOpenLanceTables = "import lancedb\n\ndb = lancedb.connect(\"hf://datasets/lance-format/pusht-lance\")\nframes = db.open_table(\"frames\")\nvideos = db.open_table(\"videos\")\n\nprint(frames.count_rows(), videos.count_rows())\nprint(frames.schema)\n";
export const PyFrameworksLerobotTrainReady = "import torch\n\ndataset = LeRobotDataset(\n \"lance-format/pusht-lance\",\n delta_timestamps={\"observation.image\": [-0.2, -0.1, 0.0]},\n return_uint8=True,\n)\nloader = torch.utils.data.DataLoader(dataset, batch_size=32, shuffle=True)\nbatch = next(iter(loader))\nprint(batch[\"observation.image\"].shape) # [32, 3, C, H, W]\n";
export const PyFrameworksLlamaindexAddReranker = "from lancedb.rerankers import ColbertReranker\n\nreranker = ColbertReranker()\nvector_store._add_reranker(reranker)\n";
export const PyFrameworksLlamaindexFiltering = "from llama_index.core.vector_stores import (\n FilterCondition,\n FilterOperator,\n MetadataFilter,\n MetadataFilters,\n)\n\nquery_filters = MetadataFilters(\n filters=[\n MetadataFilter(\n key=\"creation_date\", operator=FilterOperator.EQ, value=\"2024-05-23\"\n ),\n MetadataFilter(key=\"file_size\", value=75040, operator=FilterOperator.GT),\n ],\n condition=FilterCondition.AND,\n)\n";
export const PyFrameworksLlamaindexHybridSearch = "from lancedb.rerankers import ColbertReranker\n\nreranker = ColbertReranker()\nvector_store._add_reranker(reranker)\n\nquery_engine = index.as_query_engine(\n filters=query_filters,\n vector_store_kwargs={\n \"query_type\": \"hybrid\",\n },\n)\n\nresponse = query_engine.query(\"How much did Viaweb charge per month?\")\n";
export const PyFrameworksLlamaindexQuickStart = "import logging\nimport sys\nimport textwrap\n\nimport openai\n\n# Uncomment to see debug logs\n# logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)\n# logging.getLogger().addHandler(logging.StreamHandler(stream=sys.stdout))\nfrom llama_index.core import (\n Document,\n SimpleDirectoryReader,\n StorageContext,\n VectorStoreIndex,\n)\nfrom llama_index.vector_stores.lancedb import LanceDBVectorStore\n\nopenai.api_key = \"sk-...\"\n\ndocuments = SimpleDirectoryReader(\"./data/your-data-dir/\").load_data()\nprint(\"Document ID:\", documents[0].doc_id, \"Document Hash:\", documents[0].hash)\n\n## For LanceDB Enterprise :\n# vector_store = LanceDBVectorStore(\n# uri=\"db://db_name\", # your remote DB URI\n# api_key=\"sk_..\", # lancedb enterprise api key\n# region=\"your-region\" # the region you configured\n# host_override=\"https://your-host.com\" # if you have a custom host, otherwise omit this\n# )\n\nvector_store = LanceDBVectorStore(\n uri=\"./lancedb\", mode=\"overwrite\", query_type=\"vector\"\n)\nstorage_context = StorageContext.from_defaults(vector_store=vector_store)\n\nindex = VectorStoreIndex.from_documents(documents, storage_context=storage_context)\nlance_filter = \"metadata.file_name = 'paul_graham_essay.txt' \"\nretriever = index.as_retriever(vector_store_kwargs={\"where\": lance_filter})\nresponse = retriever.retrieve(\"What did the author do growing up?\")\n";
export const PyFrameworksPydanticBaseExample = "table = db.create_table(\"docs\", schema=LanceDocs, mode=\"overwrite\")\ntable.add(\n [\n {\"text\": \"hello world\", \"vector\": [1.0, 0.0]},\n {\"text\": \"goodbye world\", \"vector\": [0.0, 1.0]},\n ]\n)\nresults = table.search([1.0, 0.0]).limit(1).to_pydantic(LanceDocs)\nprint(results[0].text)\n# hello world\n";
export const PyFrameworksPydanticBaseModel = "class LanceDocs(LanceModel):\n text: str\n vector: Vector(2)\n";
export const PyFrameworksPydanticImports = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nfrom lancedb.pydantic import LanceModel, Vector\n";
export const PyFrameworksPydanticSetUrl = "db = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"pydantic-docs\"))\n";
export const PyFrameworksPydanticTypeConversion = "from typing import List, Optional\n\nimport pyarrow as pa\nimport pydantic\nfrom lancedb.pydantic import Vector, pydantic_to_schema\n\nclass FooModel(pydantic.BaseModel):\n id: int\n s: str\n vec: Vector(1536) # fixed_size_list<item: float32>[1536]\n li: List[int]\n\nschema = pydantic_to_schema(FooModel)\nassert schema == pa.schema(\n [\n pa.field(\"id\", pa.int64(), False),\n pa.field(\"s\", pa.utf8(), False),\n pa.field(\"vec\", pa.list_(pa.float32(), 1536)),\n pa.field(\"li\", pa.list_(pa.int64()), False),\n ]\n)\n";
export const PyFrameworksPydanticVectorField = "import pyarrow as pa\nimport pydantic\nfrom lancedb.pydantic import Vector, pydantic_to_schema\n\nclass MyModel(pydantic.BaseModel):\n id: int\n url: str\n embeddings: Vector(768)\n\nschema = pydantic_to_schema(MyModel)\nassert schema == pa.schema(\n [\n pa.field(\"id\", pa.int64(), False),\n pa.field(\"url\", pa.utf8(), False),\n pa.field(\"embeddings\", pa.list_(pa.float32(), 768)),\n ]\n)\n";
export const PyFrameworksStableWorldmodelCollectLance = "import stable_worldmodel as swm\n\nworld = swm.World(\"swm/PushT-v1\", num_envs=8)\nworld.set_policy(your_expert_policy)\nworld.collect(\"data/pusht_demo.lance\", episodes=100, seed=0)\n";
export const PyFrameworksStableWorldmodelConvert = "swm.data.convert(\n \"data/pusht_demo.lance\",\n \"data/pusht_video\",\n dest_format=\"video\",\n fps=30,\n)\n";
export const PyFrameworksStableWorldmodelEvaluate = "from stable_worldmodel.policy import PlanConfig, WorldModelPolicy\nfrom stable_worldmodel.solver import CEMSolver\n\nsolver = CEMSolver(model=world_model, num_samples=300)\npolicy = WorldModelPolicy(solver=solver, config=PlanConfig(horizon=10))\n\nworld.set_policy(policy)\nresults = world.evaluate(episodes=50)\nprint(f\"Success Rate: {results['success_rate']:.1f}%\")\n";
export const PyFrameworksStableWorldmodelLoadLance = "dataset = swm.data.load_dataset(\"data/pusht_demo.lance\", num_steps=16)\n\nbatch = dataset[0]\nprint(batch.keys())\n";
export const PyPlatformsDltAdapterImport = "from dlt.destinations.adapters import lancedb_adapter\n";
export const PyPlatformsDltAdapterUsage = "load_info = pipeline.run(\n lancedb_adapter(\n movies_source,\n embed=\"Title\",\n )\n)\n";
export const PyPlatformsDltPipeline = "# Import necessary modules\nimport dlt\nfrom rest_api import rest_api_source\n\n# Configure the REST API source\nmovies_source = rest_api_source(\n {\n \"client\": {\n \"base_url\": \"https://www.omdbapi.com/\",\n \"auth\": { # authentication strategy for the OMDb API\n \"type\": \"api_key\",\n \"name\": \"apikey\",\n \"api_key\": dlt.secrets[\n \"sources.rest_api.api_token\"\n ], # read API credentials directly from secrets.toml\n \"location\": \"query\",\n },\n \"paginator\": { # pagination strategy for the OMDb API\n \"type\": \"page_number\",\n \"base_page\": 1,\n \"total_path\": \"totalResults\",\n \"maximum_page\": 5,\n },\n },\n \"resources\": [ # list of API endpoints to request\n {\n \"name\": \"movie_search\",\n \"endpoint\": {\n \"path\": \"/\",\n \"params\": {\n \"s\": \"godzilla\",\n \"type\": \"movie\",\n },\n },\n }\n ],\n }\n)\n\nif __name__ == \"__main__\":\n # Create a pipeline object\n pipeline = dlt.pipeline(\n pipeline_name=\"movies_pipeline\",\n destination=\"lancedb\", # this tells dlt to load the data into LanceDB\n dataset_name=\"movies_data_pipeline\",\n )\n\n # Run the pipeline\n load_info = pipeline.run(movies_source)\n\n # pretty print the information on data that was loaded\n print(load_info)\n";
export const PyPlatformsDuckdbCreateTable = "import lancedb\n\ndb = lancedb.connect(\"data/sample-lancedb\")\ndata = [\n {\"vector\": [3.1, 4.1], \"item\": \"foo\", \"price\": 10.0},\n {\"vector\": [5.9, 26.5], \"item\": \"bar\", \"price\": 20.0},\n]\ntable = db.create_table(\"pd_table\", data=data)\n";
export const PyPlatformsDuckdbMeanPrice = "duckdb.query(\"SELECT mean(price) FROM arrow_table\")\n";
export const PyPlatformsDuckdbQueryTable = "import duckdb\n\narrow_table = table.to_lance()\n\nduckdb.query(\"SELECT * FROM arrow_table\")\n";
export const PyPlatformsPandasAsyncExample = "async def run_pandas_async_example() -> pd.DataFrame:\n async_db = await lancedb.connect_async(\n str(Path(tempfile.mkdtemp()) / \"pandas-async\")\n )\n async_df = pd.DataFrame(\n [\n {\"id\": \"10\", \"text\": \"sage\", \"vector\": [0.6, 0.4, 0.8]},\n {\"id\": \"11\", \"text\": \"bard\", \"vector\": [0.2, 0.7, 0.3]},\n ]\n )\n async_table = await async_db.create_table(\n \"creatures_async\", data=async_df, mode=\"overwrite\"\n )\n # An async search returns its query builder from an await\n query = await async_table.search([0.6, 0.4, 0.8])\n async_results = await query.select([\"text\", \"_distance\"]).limit(1).to_pandas()\n print(async_results)\n return async_results\n\nasync_results = asyncio.run(run_pandas_async_example())\n";
export const PyPlatformsPandasCreateTable = "pandas_df = pd.DataFrame(\n [\n {\"id\": \"1\", \"text\": \"dragon\", \"vector\": [0.9, 0.1, 0.3]},\n {\"id\": \"2\", \"text\": \"griffin\", \"vector\": [0.4, 0.5, 0.2]},\n {\"id\": \"3\", \"text\": \"phoenix\", \"vector\": [0.7, 0.3, 0.6]},\n ]\n)\npandas_db = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"pandas-demo\"))\npandas_table = pandas_db.create_table(\"creatures\", data=pandas_df, mode=\"overwrite\")\n";
export const PyPlatformsPandasImports = "import asyncio\nimport tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport pandas as pd\n";
export const PyPlatformsPandasVectorSearch = "pandas_results = (\n pandas_table.search([0.9, 0.1, 0.3])\n .select([\"text\", \"_distance\"])\n .limit(1)\n .to_pandas()\n)\nprint(pandas_results)\n";
export const PyPlatformsPolarsCreateTable = "birds = pl.DataFrame(\n {\n \"text\": [\"phoenix\", \"sparrow\"],\n \"vector\": [\n [0.1, 0.2, 0.3],\n [0.8, 0.6, 0.5],\n ],\n }\n)\npolars_db = lancedb.connect(str(Path(tempfile.mkdtemp()) / \"polars-demo\"))\npolars_table = polars_db.create_table(\n \"birds\", data=birds.to_arrow(), mode=\"overwrite\"\n)\n";
export const PyPlatformsPolarsImports = "import tempfile\nfrom pathlib import Path\n\nimport lancedb\nimport polars as pl\nfrom lancedb.pydantic import LanceModel, Vector\n";
export const PyPlatformsPolarsLazyframe = "lazy_frame = polars_table.to_polars().lazy()\nprint(lazy_frame.select([\"text\"]).collect())\n";
export const PyPlatformsPolarsPydantic = "class BirdModel(LanceModel):\n text: str\n vector: Vector(3)\n\nschema_table = polars_db.create_table(\n \"birds_schema\", schema=BirdModel, mode=\"overwrite\"\n)\nschema_table.add(birds.to_dicts())\n";
export const PyPlatformsPolarsVectorSearch = "polars_results = (\n polars_table.search([0.1, 0.2, 0.3])\n .select([\"text\", \"_distance\"])\n .limit(1)\n .to_polars()\n)\nprint(polars_results)\n";
export const PyPlatformsVoxel51BackendFlag = "import fiftyone.brain as fob\n\n# Re-run similarity creation using the LanceDB backend explicitly\nfob.compute_similarity(\n dataset,\n model=\"clip-vit-base32-torch\",\n brain_key=\"lancedb_index\",\n backend=\"lancedb\",\n)\n";
export const PyPlatformsVoxel51BackendParams = "lancedb_index = fob.compute_similarity(\n dataset,\n model=\"clip-vit-base32-torch\",\n backend=\"lancedb\",\n brain_key=\"lancedb_index\",\n table_name=\"your-table\",\n metric=\"euclidean\",\n uri=\"/tmp/lancedb\",\n)\n";
export const PyPlatformsVoxel51BrainConfig = "import fiftyone.brain as fob\n\n# Print your current brain config\nprint(fob.brain_config)\n";
export const PyPlatformsVoxel51Cleanup = "# Step 5 (optional): Cleanup\n\n# Delete the LanceDB table\nlancedb_index.cleanup()\n\n# Delete run record from FiftyOne\ndataset.delete_brain_run(\"lancedb_index\")\n";
export const PyPlatformsVoxel51ComputeSimilarity = "# Steps 2 and 3: Compute embeddings and create a similarity index\nlancedb_index = fob.compute_similarity(\n dataset,\n model=\"clip-vit-base32-torch\",\n brain_key=\"lancedb_index\",\n backend=\"lancedb\",\n)\n";
export const PyPlatformsVoxel51LoadDataset = "import fiftyone as fo\nimport fiftyone.brain as fob\nimport fiftyone.zoo as foz\n\n# Step 1: Load your data into FiftyOne\ndataset = foz.load_zoo_dataset(\"quickstart\")\n";
export const PyPlatformsVoxel51SortBySimilarity = "# Step 4: Query your data\nquery = dataset.first().id # query by sample ID\nview = dataset.sort_by_similarity(\n query,\n brain_key=\"lancedb_index\",\n k=10, # limit to 10 most similar samples\n)\n";
export const PyRerankingAnswerdotaiUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import AnswerdotaiRerankers\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = AnswerdotaiRerankers()\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingCohereUsage = "import os\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import CohereReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = CohereReranker(api_key=os.environ[\"COHERE_API_KEY\"])\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingColbertUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import ColbertReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = ColbertReranker()\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingCrossEncoderUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import CrossEncoderReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = CrossEncoderReranker()\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingJinaUsage = "import os\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import JinaReranker\n\nembedder = get_registry().get(\"jina\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = JinaReranker(api_key=os.environ[\"JINA_API_KEY\"])\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingLinearCombinationUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import LinearCombinationReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = LinearCombinationReranker()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingMrrUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import MRRReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = MRRReranker(weight_vector=0.7, weight_fts=0.3)\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n\n# Run multivector search across multiple vector columns\nrs1 = tbl.search(\"hello\").limit(10).with_row_id(True).to_arrow()\nrs2 = tbl.search(\"greeting\").limit(10).with_row_id(True).to_arrow()\ncombined = MRRReranker().rerank_multivector([rs1, rs2])\n";
export const PyRerankingOpenaiUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import OpenaiReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = OpenaiReranker()\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingRrfUsage = "import lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import RRFReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = RRFReranker()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingVoyageaiUsage = "import os\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import VoyageAIReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\nreranker = VoyageAIReranker(model_name=\"rerank-2\")\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const PyRerankingWatsonxUsage = "import os\n\nimport lancedb\nfrom lancedb.embeddings import get_registry\nfrom lancedb.pydantic import LanceModel, Vector\nfrom lancedb.rerankers import WatsonxReranker\n\nembedder = get_registry().get(\"sentence-transformers\").create()\ndb = lancedb.connect(\"~/.lancedb\")\n\nclass Schema(LanceModel):\n text: str = embedder.SourceField()\n vector: Vector(embedder.ndims()) = embedder.VectorField()\n\ndata = [\n {\"text\": \"hello world\"},\n {\"text\": \"goodbye world\"},\n]\ntbl = db.create_table(\"test\", schema=Schema, mode=\"overwrite\")\ntbl.add(data)\n\n# Credentials pulled from WATSONX_API_KEY and WATSONX_PROJECT_ID (or WATSONX_SPACE_ID)\nreranker = WatsonxReranker(\n api_key=os.environ[\"WATSONX_API_KEY\"],\n project_id=os.environ[\"WATSONX_PROJECT_ID\"],\n)\n\n# Run vector search with a reranker\nresult = tbl.search(\"hello\").rerank(reranker=reranker).to_list()\n\n# Run FTS search with a reranker\nresult = tbl.search(\"hello\", query_type=\"fts\").rerank(reranker=reranker).to_list()\n\n# Run hybrid search with a reranker\ntbl.create_fts_index(\"text\", replace=True)\nresult = (\n tbl.search(\"hello\", query_type=\"hybrid\").rerank(reranker=reranker).to_list()\n)\n";
export const TsFrameworksGenkitCustomIndexer = "export const menuPdfIndexer = lancedbIndexerRef({\n // Using all defaults, for dbUri, tableName, and embedder, etc\n});\n\nconst chunkingConfig = {\n minLength: 1000,\n maxLength: 2000,\n splitter: \"sentence\",\n overlap: 100,\n delimiters: \"\",\n} as any;\n\nasync function extractTextFromPdf(filePath: string) {\n const pdfFile = path.resolve(filePath);\n const dataBuffer = await readFile(pdfFile);\n const data = await pdf(dataBuffer);\n return data.text;\n}\n\nexport const indexMenu = ai.defineFlow(\n {\n name: \"indexMenu\",\n inputSchema: z.string().describe(\"PDF file path\"),\n outputSchema: z.void(),\n },\n async (filePath: string) => {\n filePath = path.resolve(filePath);\n\n // Read the pdf.\n const pdfTxt = await ai.run(\"extract-text\", () => extractTextFromPdf(filePath));\n\n // Divide the pdf text into segments.\n const chunks = await ai.run(\"chunk-it\", async () => chunk(pdfTxt, chunkingConfig));\n\n // Convert chunks of text into documents to store in the index.\n const documents = chunks.map((text) => {\n return Document.fromText(text, { filePath });\n });\n\n // Add documents to the index.\n await ai.index({\n indexer: menuPdfIndexer,\n documents,\n options: {\n writeMode: WriteMode.Overwrite,\n } as any,\n });\n },\n);\n";
export const TsFrameworksGenkitCustomRetriever = "export const menuRetriever = lancedbRetrieverRef({\n tableName: \"table\", // Use the same table name as the indexer.\n displayName: \"Menu\", // Use a custom display name.\n});\n\nexport const menuQAFlow = ai.defineFlow(\n { name: \"Menu\", inputSchema: z.string(), outputSchema: z.string() },\n async (input: string) => {\n // retrieve relevant documents\n const docs = await ai.retrieve({\n retriever: menuRetriever,\n query: input,\n options: {\n k: 3,\n },\n });\n\n const extractedContent = docs.map((doc) => {\n if (doc.content && Array.isArray(doc.content) && doc.content.length > 0) {\n if (doc.content[0].media && doc.content[0].media.url) {\n return doc.content[0].media.url;\n }\n }\n return \"No content found\";\n });\n\n console.log(\"Extracted content:\", extractedContent);\n\n const { text } = await ai.generate({\n model: gemini(\"gemini-2.0-flash\"),\n prompt: `\nYou are acting as a helpful AI assistant that can answer \nquestions about the food available on the menu at Genkit Grub Pub.\n\nUse only the context provided to answer the question.\nIf you don't know, do not make up an answer.\nDo not add or change items on the menu.\n\nContext:\n${extractedContent.join(\"\\n\\n\")}\n\nQuestion: ${input}`,\n docs,\n });\n\n return text;\n },\n);\n";
export const TsFrameworksGenkitUsage = "import { lancedbIndexerRef, lancedb, lancedbRetrieverRef, WriteMode } from \"genkitx-lancedb\";\nimport { textEmbedding004, vertexAI } from \"@genkit-ai/vertexai\";\nimport { gemini } from \"@genkit-ai/vertexai\";\nimport { z, genkit } from \"genkit\";\nimport { Document } from \"genkit/retriever\";\nimport { chunk } from \"llm-chunk\";\nimport { readFile } from \"fs/promises\";\nimport path from \"path\";\nimport pdf from \"pdf-parse/lib/pdf-parse\";\n\nconst ai = genkit({\n plugins: [\n // vertexAI provides the textEmbedding004 embedder\n vertexAI(),\n\n // the local vector store requires an embedder to translate from text to vector\n lancedb([\n {\n dbUri: \".db\", // optional lancedb uri, default to .db\n tableName: \"table\", // optional table name, default to table\n embedder: textEmbedding004,\n },\n ]),\n ],\n});\n";
Reference in New Issue View Git Blame Copy Permalink