The AI Toolchain — issue -415, April 27, 2021

Haystack v0.8.0 adds MilvusDocumentStore, Knowledge Graph QA, YAML Pipeline config, confidence scores, and a Selenium web crawler.

$ git clone --branch v0.8.0 https://github.com/deepset-ai/haystack.git
# already have the repo? check out this version:
$ git checkout v0.8.0

$ curl -X POST http://localhost:8000/query \
  -H 'Content-Type: application/json' \
  -d '{"query": "Why did the revenue change?"}'

• ›Adds MilvusDocumentStore class enabling embedding-based retrievers (DensePassageRetriever, EmbeddingRetriever) to use production-ready Milvus vector database servers for large-scale deployments.
• ›Adds GraphDBKnowlegeGraph class for storing RDF Triples and executing SPARQL queries, integrable with the new Text2SparqlRetriever to convert natural language queries to SPARQL.
• ›Introduces YAML-based Pipeline configuration via rest_api/pipeline.yaml, enabling shareable query and indexing configs, reproducible setups, and A/B testing of Pipelines.
• ›Adds new generic POST /query endpoint to the REST API backed by Pipelines, replacing the former /doc-qa and /faq-qa endpoints; accepts a single query string and returns answers with a probability confidence score (range 0–1).
• ›Adds new generic POST /feedback endpoint, replacing the former /doc-qa-feedback and /faq-qa-feedback endpoints.

Triton v2.9.0 adds GKE Marketplace deployment, GRPC compression, ragged batching for TensorFlow, and ensemble unload control.

$ git clone --branch v2.9.0 https://github.com/triton-inference-server/server.git
# already have the repo? check out this version:
$ git checkout v2.9.0

parameters {
  key: "TF_GRAPH_TAG"
  value: { string_value: "serve" }
}
parameters {
  key: "TF_SIGNATURE_DEF"
  value: { string_value: "serving_default" }
}

• ›The Model Control API gains an option to unload all models contained within an ensemble when unloading the ensemble itself — see the model-repository extension protocol docs.
• ›GRPC client libraries now support enabling compression.
• ›Ragged batching is now supported for TensorFlow models.
• ›For TensorFlow SavedModel-format models, adds the ability to select which graph and signature_def to load via backend parameters.
• ›Triton Server is now available as a GKE Marketplace Application.

Qdrant v0.2.0 adds payload field indexing for number and keyword types and revamps the Search API filter schema.

$ git clone --branch v0.2.0 https://github.com/qdrant/qdrant.git
# already have the repo? check out this version:
$ git checkout v0.2.0

$ curl -X POST 'http://localhost:6333/collections/{collection_name}/points/search' \
  -H 'Content-Type: application/json' \
  -d '{
    "vector": [0.1, 0.2, 0.3],
    "filter": {
      "should": [
        {
          "key": "city",
          "match": {
            "keyword": "London"
          }
        }
      ]
    },
    "top": 5
  }'

• ›Adds the ability to mark a payload field as 'indexed', enabling payload indexation for number and keyword field types (geo field index is in progress).
• ›Segments are now automatically rebuilt into MMap and indexed segments when enough vectors are stored, unlocking better performance at scale without manual intervention.
• ›Changes the Search API filter schema: the key property moves out of the condition object (e.g. match) and becomes a sibling field alongside it — existing filter queries using the old nested-key layout will break.

Weaviate v1.3.0 adds a BERT-based Q&A module with a new ask{} GraphQL searcher and richer transformer model metadata via /v1/meta.

$ git clone --branch v1.3.0 https://github.com/weaviate/weaviate.git
# already have the repo? check out this version:
$ git checkout v1.3.0

{
  Get {
    Paragraph(
      ask: {
        question: "what is the population of Berlin?"
        certainty: 0.8
      }
    ) {
      _additional { answer { hasAnswer result certainty property startPosition endPosition } }
      text
    }
  }
}

• ›Introduces the qna-transformers module, enabling BERT-style answer extraction via a new ask{} searcher on GraphQL Get { ... } queries, configured with a "question" (required string), optional "certainty" (float 0..1), and optional "properties" ([]string).
• ›Adds a new _additional { answer { } } response field containing hasAnswer (boolean), result (nullable string), certainty (nullable float), property (nullable string), startPosition (int), and endPosition (int) — surfacing extracted answers directly in query results.
• ›Supports custom Hugging Face models for Q&A via the semitechnologies/qna-transformers:custom base image, compatible with transformers.AutoModelForQuestionAnswering.
• ›Expands the GET /v1/meta endpoint to include meta information about transformer models in use across all transformer-based modules.