Datasets:

vblagoje

/

haystack-pipelines

like 1

Build Haystack question answer generation pipeline using TableReader and question generator

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "table_reader", "type": "TableReader", "params": {"model_name_or_path": "google/tapas-base-finetuned-wtq", "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answer": false, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "table_reader"}]}]}

Build search pipeline using PineconeDocumentStore and ElasticsearchRetriever

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "validate_index_sync": true}}, {"name": "elasticsearch_retriever", "type": "ElasticsearchRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_retriever"}, {"inputs": ["elasticsearch_retriever"]}]}]}

Make faq system consisting of WeaviateDocumentStore and table text retriever

{"version": "1.8.0", "components": [{"name": "weaviate_document_store", "type": "WeaviateDocumentStore", "params": {"port": 0, "timeout_config": [5, 15], "index": "Document", "embedding_dim": 768, "content_field": "content", "name_field": "name", "similarity": "cosine", "index_type": "hnsw", "return_embedding": false, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"]}]}]}

Build extractive qa pipeline using bm25 retriever, pinecone document store and rci reader

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "metadata_config": {}, "validate_index_sync": true}}, {"name": "bm25_retriever", "type": "BM25Retriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}, {"name": "rci_reader", "type": "RCIReader", "params": {"row_model_name_or_path": "michaelrglass/albert-base-rci-wikisql-row", "column_model_name_or_path": "michaelrglass/albert-base-rci-wikisql-col", "use_gpu": true, "top_k": 10, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "bm25_retriever"}, {"inputs": ["bm25_retriever"], "name": "rci_reader"}]}]}

Create QuestionGenerationPipeline

{"version": "1.8.0", "components": [{"name": "QuestionGenerator", "params": {}, "type": "QuestionGenerator"}], "pipelines": [{"name": "query", "nodes": [{"inputs": ["Query"], "name": "QuestionGenerator"}]}]}

Build Haystack qa system with SQLDocumentStore, table reader and TableTextRetriever

{"version": "1.8.0", "components": [{"name": "sql_document_store", "type": "SQLDocumentStore", "params": {"url": "sqlite://", "index": "document", "label_index": "label", "duplicate_documents": "overwrite", "check_same_thread": false}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}, {"name": "table_reader", "type": "TableReader", "params": {"model_name_or_path": "google/tapas-base-finetuned-wtq", "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answer": false, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"], "name": "table_reader"}]}]}

Generate generative pipeline with seq2 seq generator, ElasticsearchRetriever and OpenDistroElasticsearchDocumentStore

{"version": "1.8.0", "components": [{"name": "open_distro_elasticsearch_document_store", "type": "OpenDistroElasticsearchDocumentStore", "params": {"scheme": "https", "username": "admin", "password": "admin", "port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "verify_certs": false, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "cosine", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false}}, {"name": "elasticsearch_retriever", "type": "ElasticsearchRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}, {"name": "seq2_seq_generator", "type": "Seq2SeqGenerator", "params": {"top_k": 1, "max_length": 200, "min_length": 2, "num_beams": 8, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_retriever"}, {"inputs": ["elasticsearch_retriever"], "name": "seq2_seq_generator"}]}]}

Make Haystack ExtractiveQAPipeline using TfidfRetriever, faiss document store and FARMReader

{"version": "1.8.0", "components": [{"name": "faiss_document_store", "type": "FAISSDocumentStore", "params": {"sql_url": "sqlite:///faiss_document_store.db", "vector_dim": 0, "embedding_dim": 768, "faiss_index_factory_str": "Flat", "return_embedding": false, "index": "document", "similarity": "dot_product", "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "n_links": 64, "ef_search": 20, "ef_construction": 80, "validate_index_sync": true}}, {"name": "tfidf_retriever", "type": "TfidfRetriever", "params": {"top_k": 10}}, {"name": "farm_reader", "type": "FARMReader", "params": {"context_window_size": 150, "batch_size": 50, "use_gpu": true, "no_ans_boost": 0.0, "return_no_answer": false, "top_k": 10, "top_k_per_candidate": 3, "top_k_per_sample": 1, "num_processes": 0, "max_seq_len": 256, "doc_stride": 128, "progress_bar": true, "duplicate_filtering": 0, "use_confidence_scores": true, "confidence_threshold": 0, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "tfidf_retriever"}, {"inputs": ["tfidf_retriever"], "name": "farm_reader"}]}]}

Build QuestionAnswerGenerationPipeline with question generator and TransformersReader

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "transformers_reader", "type": "TransformersReader", "params": {"model_name_or_path": "distilbert-base-uncased-distilled-squad", "context_window_size": 70, "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answers": false, "max_seq_len": 256, "doc_stride": 128, "batch_size": 16, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "transformers_reader"}]}]}

Generate GenerativeQAPipeline using tfidf retriever, InMemoryDocumentStore and ra generator

{"version": "1.8.0", "components": [{"name": "in_memory_document_store", "type": "InMemoryDocumentStore", "params": {"index": "document", "label_index": "label", "embedding_dim": 768, "return_embedding": false, "similarity": "dot_product", "progress_bar": true, "duplicate_documents": "overwrite", "use_gpu": true, "scoring_batch_size": 500000}}, {"name": "tfidf_retriever", "type": "TfidfRetriever", "params": {"top_k": 10}}, {"name": "ra_generator", "type": "RAGenerator", "params": {"model_name_or_path": "facebook/rag-token-nq", "generator_type": "token", "top_k": 2, "max_length": 200, "min_length": 2, "num_beams": 2, "embed_title": true, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "tfidf_retriever"}, {"inputs": ["tfidf_retriever"], "name": "ra_generator"}]}]}

Make document search pipeline with in memory document store and ElasticsearchRetriever

{"version": "1.8.0", "components": [{"name": "in_memory_document_store", "type": "InMemoryDocumentStore", "params": {"index": "document", "label_index": "label", "embedding_dim": 768, "return_embedding": false, "similarity": "dot_product", "progress_bar": true, "duplicate_documents": "overwrite", "use_gpu": true, "scoring_batch_size": 500000}}, {"name": "elasticsearch_retriever", "type": "ElasticsearchRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_retriever"}, {"inputs": ["elasticsearch_retriever"]}]}]}

Generate GenerativeQAPipeline consisting of SQLDocumentStore, seq2 seq generator and DensePassageRetriever

{"version": "1.8.0", "components": [{"name": "sql_document_store", "type": "SQLDocumentStore", "params": {"url": "sqlite://", "index": "document", "label_index": "label", "duplicate_documents": "overwrite", "check_same_thread": false}}, {"name": "dense_passage_retriever", "type": "DensePassageRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_title": true, "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true}}, {"name": "seq2_seq_generator", "type": "Seq2SeqGenerator", "params": {"top_k": 1, "max_length": 200, "min_length": 2, "num_beams": 8, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "dense_passage_retriever"}, {"inputs": ["dense_passage_retriever"], "name": "seq2_seq_generator"}]}]}

Build Haystack GenerativeQAPipeline with PineconeDocumentStore, TfidfRetriever and open ai answer generator

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "validate_index_sync": true}}, {"name": "tfidf_retriever", "type": "TfidfRetriever", "params": {"top_k": 10}}, {"name": "open_ai_answer_generator", "type": "OpenAIAnswerGenerator", "params": {"model": "text-curie-001", "max_tokens": 7, "top_k": 5, "temperature": 0, "presence_penalty": -2.0, "frequency_penalty": -2.0, "progress_bar": true}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "tfidf_retriever"}, {"inputs": ["tfidf_retriever"], "name": "open_ai_answer_generator"}]}]}

Create FAQPipeline using FAISSDocumentStore and bm25 retriever

{"version": "1.8.0", "components": [{"name": "faiss_document_store", "type": "FAISSDocumentStore", "params": {"sql_url": "sqlite:///faiss_document_store.db", "vector_dim": 0, "embedding_dim": 768, "faiss_index_factory_str": "Flat", "return_embedding": false, "index": "document", "similarity": "dot_product", "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "n_links": 64, "ef_search": 20, "ef_construction": 80, "validate_index_sync": true}}, {"name": "bm25_retriever", "type": "BM25Retriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "bm25_retriever"}, {"inputs": ["bm25_retriever"]}]}]}

Generate Haystack generative qa pipeline using seq2 seq generator, SQLDocumentStore and TableTextRetriever

{"version": "1.8.0", "components": [{"name": "sql_document_store", "type": "SQLDocumentStore", "params": {"url": "sqlite://", "index": "document", "label_index": "label", "duplicate_documents": "overwrite", "check_same_thread": false}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}, {"name": "seq2_seq_generator", "type": "Seq2SeqGenerator", "params": {"top_k": 1, "max_length": 200, "min_length": 2, "num_beams": 8, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"], "name": "seq2_seq_generator"}]}]}

Generate faq pipeline using EmbeddingRetriever and OpenDistroElasticsearchDocumentStore

{"version": "1.8.0", "components": [{"name": "open_distro_elasticsearch_document_store", "type": "OpenDistroElasticsearchDocumentStore", "params": {"scheme": "https", "username": "admin", "password": "admin", "port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "verify_certs": false, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "cosine", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false}}, {"name": "embedding_retriever", "type": "EmbeddingRetriever", "params": {"use_gpu": true, "batch_size": 32, "max_seq_len": 512, "pooling_strategy": "reduce_mean", "emb_extraction_layer": -1, "top_k": 10, "progress_bar": true, "use_auth_token": false, "scale_score": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "embedding_retriever"}, {"inputs": ["embedding_retriever"]}]}]}

Generate Haystack QuestionGenerationPipeline

{"version": "1.8.0", "components": [{"name": "QuestionGenerator", "params": {}, "type": "QuestionGenerator"}], "pipelines": [{"name": "query", "nodes": [{"inputs": ["Query"], "name": "QuestionGenerator"}]}]}

Make Haystack FAQPipeline with WeaviateDocumentStore and MultihopEmbeddingRetriever

{"version": "1.8.0", "components": [{"name": "weaviate_document_store", "type": "WeaviateDocumentStore", "params": {"port": 0, "timeout_config": [5, 15], "index": "Document", "embedding_dim": 768, "content_field": "content", "name_field": "name", "similarity": "cosine", "index_type": "hnsw", "return_embedding": false, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false}}, {"name": "multihop_embedding_retriever", "type": "MultihopEmbeddingRetriever", "params": {"num_iterations": 2, "use_gpu": true, "batch_size": 32, "max_seq_len": 512, "model_format": "farm", "pooling_strategy": "reduce_mean", "emb_extraction_layer": -1, "top_k": 10, "progress_bar": true, "use_auth_token": false, "scale_score": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "multihop_embedding_retriever"}, {"inputs": ["multihop_embedding_retriever"]}]}]}

Build FAQPipeline using open search document store and table text retriever

{"version": "1.8.0", "components": [{"name": "open_search_document_store", "type": "OpenSearchDocumentStore", "params": {"scheme": "https", "username": "admin", "password": "admin", "port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "verify_certs": false, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "dot_product", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false, "knn_engine": "nmslib"}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"]}]}]}

Generate FAQPipeline using DeepsetCloudDocumentStore and ElasticsearchRetriever

{"version": "1.8.0", "components": [{"name": "deepset_cloud_document_store", "type": "DeepsetCloudDocumentStore", "params": {"workspace": "default", "duplicate_documents": "overwrite", "similarity": "dot_product", "return_embedding": false, "label_index": "default", "embedding_dim": 768}}, {"name": "elasticsearch_retriever", "type": "ElasticsearchRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_retriever"}, {"inputs": ["elasticsearch_retriever"]}]}]}

Create Haystack generative pipeline consisting of embedding retriever, Seq2SeqGenerator and DeepsetCloudDocumentStore

{"version": "1.8.0", "components": [{"name": "deepset_cloud_document_store", "type": "DeepsetCloudDocumentStore", "params": {"workspace": "default", "duplicate_documents": "overwrite", "similarity": "dot_product", "return_embedding": false, "label_index": "default", "embedding_dim": 768}}, {"name": "embedding_retriever", "type": "EmbeddingRetriever", "params": {"use_gpu": true, "batch_size": 32, "max_seq_len": 512, "pooling_strategy": "reduce_mean", "emb_extraction_layer": -1, "top_k": 10, "progress_bar": true, "use_auth_token": false, "scale_score": true}}, {"name": "seq2_seq_generator", "type": "Seq2SeqGenerator", "params": {"top_k": 1, "max_length": 200, "min_length": 2, "num_beams": 8, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "embedding_retriever"}, {"inputs": ["embedding_retriever"], "name": "seq2_seq_generator"}]}]}

Generate qa pipeline using elasticsearch document store, filter retriever and rci reader

{"version": "1.8.0", "components": [{"name": "elasticsearch_document_store", "type": "ElasticsearchDocumentStore", "params": {"port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "scheme": "http", "verify_certs": true, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "dot_product", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false}}, {"name": "filter_retriever", "type": "FilterRetriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}, {"name": "rci_reader", "type": "RCIReader", "params": {"row_model_name_or_path": "michaelrglass/albert-base-rci-wikisql-row", "column_model_name_or_path": "michaelrglass/albert-base-rci-wikisql-col", "use_gpu": true, "top_k": 10, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "filter_retriever"}, {"inputs": ["filter_retriever"], "name": "rci_reader"}]}]}

Create Haystack extractive qa system using TableReader, TableTextRetriever and OpenSearchDocumentStore

{"version": "1.8.0", "components": [{"name": "open_search_document_store", "type": "OpenSearchDocumentStore", "params": {"scheme": "https", "username": "admin", "password": "admin", "port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "verify_certs": false, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "dot_product", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false, "knn_engine": "nmslib"}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}, {"name": "table_reader", "type": "TableReader", "params": {"model_name_or_path": "google/tapas-base-finetuned-wtq", "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answer": false, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"], "name": "table_reader"}]}]}

Build Haystack qa pipeline with FARMReader, WeaviateDocumentStore and MultihopEmbeddingRetriever

{"version": "1.8.0", "components": [{"name": "weaviate_document_store", "type": "WeaviateDocumentStore", "params": {"port": 0, "timeout_config": [5, 15], "index": "Document", "embedding_dim": 768, "content_field": "content", "name_field": "name", "similarity": "cosine", "index_type": "hnsw", "return_embedding": false, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false}}, {"name": "multihop_embedding_retriever", "type": "MultihopEmbeddingRetriever", "params": {"num_iterations": 2, "use_gpu": true, "batch_size": 32, "max_seq_len": 512, "model_format": "farm", "pooling_strategy": "reduce_mean", "emb_extraction_layer": -1, "top_k": 10, "progress_bar": true, "use_auth_token": false, "scale_score": true}}, {"name": "farm_reader", "type": "FARMReader", "params": {"context_window_size": 150, "batch_size": 50, "use_gpu": true, "no_ans_boost": 0.0, "return_no_answer": false, "top_k": 10, "top_k_per_candidate": 3, "top_k_per_sample": 1, "num_processes": 0, "max_seq_len": 256, "doc_stride": 128, "progress_bar": true, "duplicate_filtering": 0, "use_confidence_scores": true, "confidence_threshold": 0, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "multihop_embedding_retriever"}, {"inputs": ["multihop_embedding_retriever"], "name": "farm_reader"}]}]}

Create Haystack generative qa system with TfidfRetriever, pinecone document store and OpenAIAnswerGenerator

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "validate_index_sync": true}}, {"name": "tfidf_retriever", "type": "TfidfRetriever", "params": {"top_k": 10}}, {"name": "open_ai_answer_generator", "type": "OpenAIAnswerGenerator", "params": {"model": "text-curie-001", "max_tokens": 7, "top_k": 5, "temperature": 0, "presence_penalty": -2.0, "frequency_penalty": -2.0, "progress_bar": true}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "tfidf_retriever"}, {"inputs": ["tfidf_retriever"], "name": "open_ai_answer_generator"}]}]}

Make Haystack faq pipeline consisting of FilterRetriever and InMemoryDocumentStore

{"version": "1.8.0", "components": [{"name": "in_memory_document_store", "type": "InMemoryDocumentStore", "params": {"index": "document", "label_index": "label", "embedding_dim": 768, "return_embedding": false, "similarity": "dot_product", "progress_bar": true, "duplicate_documents": "overwrite", "use_gpu": true, "scoring_batch_size": 500000}}, {"name": "filter_retriever", "type": "FilterRetriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "filter_retriever"}, {"inputs": ["filter_retriever"]}]}]}

Generate Haystack QuestionAnswerGenerationPipeline using question generator and transformers reader

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "transformers_reader", "type": "TransformersReader", "params": {"model_name_or_path": "distilbert-base-uncased-distilled-squad", "context_window_size": 70, "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answers": false, "max_seq_len": 256, "doc_stride": 128, "batch_size": 16, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "transformers_reader"}]}]}

Build Haystack question generation pipeline

{"version": "1.8.0", "components": [{"name": "QuestionGenerator", "params": {}, "type": "QuestionGenerator"}], "pipelines": [{"name": "query", "nodes": [{"inputs": ["Query"], "name": "QuestionGenerator"}]}]}

Build question answer generation pipeline with question generator and rci reader

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "rci_reader", "type": "RCIReader", "params": {"row_model_name_or_path": "michaelrglass/albert-base-rci-wikisql-row", "column_model_name_or_path": "michaelrglass/albert-base-rci-wikisql-col", "use_gpu": true, "top_k": 10, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "rci_reader"}]}]}

Generate Haystack QuestionAnswerGenerationPipeline using TableReader and question generator

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "table_reader", "type": "TableReader", "params": {"model_name_or_path": "google/tapas-base-finetuned-wtq", "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answer": false, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "table_reader"}]}]}

Create QuestionGenerationPipeline

{"version": "1.8.0", "components": [{"name": "QuestionGenerator", "params": {}, "type": "QuestionGenerator"}], "pipelines": [{"name": "query", "nodes": [{"inputs": ["Query"], "name": "QuestionGenerator"}]}]}

Make Haystack question answer generation pipeline using TableReader and question generator

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "table_reader", "type": "TableReader", "params": {"model_name_or_path": "google/tapas-base-finetuned-wtq", "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answer": false, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "table_reader"}]}]}

Make faq system with multihop embedding retriever and weaviate document store

{"version": "1.8.0", "components": [{"name": "weaviate_document_store", "type": "WeaviateDocumentStore", "params": {"port": 0, "timeout_config": [5, 15], "index": "Document", "embedding_dim": 768, "content_field": "content", "name_field": "name", "similarity": "cosine", "index_type": "hnsw", "return_embedding": false, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false}}, {"name": "multihop_embedding_retriever", "type": "MultihopEmbeddingRetriever", "params": {"num_iterations": 2, "use_gpu": true, "batch_size": 32, "max_seq_len": 512, "model_format": "farm", "pooling_strategy": "reduce_mean", "emb_extraction_layer": -1, "top_k": 10, "progress_bar": true, "use_auth_token": false, "scale_score": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "multihop_embedding_retriever"}, {"inputs": ["multihop_embedding_retriever"]}]}]}

Make search summarization system using TransformersSummarizer, pinecone document store and bm25 retriever

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "validate_index_sync": true}}, {"name": "bm25_retriever", "type": "BM25Retriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}, {"name": "transformers_summarizer", "type": "TransformersSummarizer", "params": {"model_name_or_path": "google/pegasus-xsum", "max_length": 200, "min_length": 5, "use_gpu": true, "clean_up_tokenization_spaces": true, "separator_for_single_summary": " ", "generate_single_summary": false, "batch_size": 16, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "bm25_retriever"}, {"inputs": ["bm25_retriever"], "name": "transformers_summarizer"}]}]}

Create generative qa system using DeepsetCloudDocumentStore, OpenAIAnswerGenerator and TfidfRetriever

{"version": "1.8.0", "components": [{"name": "deepset_cloud_document_store", "type": "DeepsetCloudDocumentStore", "params": {"workspace": "default", "duplicate_documents": "overwrite", "similarity": "dot_product", "return_embedding": false, "label_index": "default", "embedding_dim": 768}}, {"name": "tfidf_retriever", "type": "TfidfRetriever", "params": {"top_k": 10}}, {"name": "open_ai_answer_generator", "type": "OpenAIAnswerGenerator", "params": {"model": "text-curie-001", "max_tokens": 7, "top_k": 5, "temperature": 0, "presence_penalty": -2.0, "frequency_penalty": -2.0, "progress_bar": true}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "tfidf_retriever"}, {"inputs": ["tfidf_retriever"], "name": "open_ai_answer_generator"}]}]}

Make Haystack question answer generation system with TransformersReader and question generator

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "transformers_reader", "type": "TransformersReader", "params": {"model_name_or_path": "distilbert-base-uncased-distilled-squad", "context_window_size": 70, "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answers": false, "max_seq_len": 256, "doc_stride": 128, "batch_size": 16, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "transformers_reader"}]}]}

Create Haystack document search system consisting of FAISSDocumentStore and elasticsearch retriever

{"version": "1.8.0", "components": [{"name": "faiss_document_store", "type": "FAISSDocumentStore", "params": {"sql_url": "sqlite:///faiss_document_store.db", "vector_dim": 0, "embedding_dim": 768, "faiss_index_factory_str": "Flat", "return_embedding": false, "index": "document", "similarity": "dot_product", "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "n_links": 64, "ef_search": 20, "ef_construction": 80, "validate_index_sync": true}}, {"name": "elasticsearch_retriever", "type": "ElasticsearchRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_retriever"}, {"inputs": ["elasticsearch_retriever"]}]}]}

Build extractive qa using transformers reader, MultihopEmbeddingRetriever and SQLDocumentStore

{"version": "1.8.0", "components": [{"name": "sql_document_store", "type": "SQLDocumentStore", "params": {"url": "sqlite://", "index": "document", "label_index": "label", "duplicate_documents": "overwrite", "check_same_thread": false}}, {"name": "multihop_embedding_retriever", "type": "MultihopEmbeddingRetriever", "params": {"num_iterations": 2, "use_gpu": true, "batch_size": 32, "max_seq_len": 512, "model_format": "farm", "pooling_strategy": "reduce_mean", "emb_extraction_layer": -1, "top_k": 10, "progress_bar": true, "use_auth_token": false, "scale_score": true}}, {"name": "transformers_reader", "type": "TransformersReader", "params": {"model_name_or_path": "distilbert-base-uncased-distilled-squad", "context_window_size": 70, "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answers": false, "max_seq_len": 256, "doc_stride": 128, "batch_size": 16, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "multihop_embedding_retriever"}, {"inputs": ["multihop_embedding_retriever"], "name": "transformers_reader"}]}]}

Create Haystack generative pipeline consisting of DensePassageRetriever, SQLDocumentStore and RAGenerator

{"version": "1.8.0", "components": [{"name": "sql_document_store", "type": "SQLDocumentStore", "params": {"url": "sqlite://", "index": "document", "label_index": "label", "duplicate_documents": "overwrite", "check_same_thread": false}}, {"name": "dense_passage_retriever", "type": "DensePassageRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_title": true, "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true}}, {"name": "ra_generator", "type": "RAGenerator", "params": {"model_name_or_path": "facebook/rag-token-nq", "generator_type": "token", "top_k": 2, "max_length": 200, "min_length": 2, "num_beams": 2, "embed_title": true, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "dense_passage_retriever"}, {"inputs": ["dense_passage_retriever"], "name": "ra_generator"}]}]}

Create Haystack generative pipeline with open search document store, FilterRetriever and ra generator

{"version": "1.8.0", "components": [{"name": "open_search_document_store", "type": "OpenSearchDocumentStore", "params": {"scheme": "https", "username": "admin", "password": "admin", "port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "verify_certs": false, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "dot_product", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false, "knn_engine": "nmslib"}}, {"name": "filter_retriever", "type": "FilterRetriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}, {"name": "ra_generator", "type": "RAGenerator", "params": {"model_name_or_path": "facebook/rag-token-nq", "generator_type": "token", "top_k": 2, "max_length": 200, "min_length": 2, "num_beams": 2, "embed_title": true, "use_gpu": true, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_generator_pipeline", "nodes": [{"inputs": ["Query"], "name": "filter_retriever"}, {"inputs": ["filter_retriever"], "name": "ra_generator"}]}]}

Make Haystack document search system consisting of TfidfRetriever and open distro elasticsearch document store

{"version": "1.8.0", "components": [{"name": "open_distro_elasticsearch_document_store", "type": "OpenDistroElasticsearchDocumentStore", "params": {"scheme": "https", "username": "admin", "password": "admin", "port": 0, "index": "document", "label_index": "label", "content_field": "content", "name_field": "name", "embedding_field": "embedding", "embedding_dim": 768, "analyzer": "standard", "verify_certs": false, "recreate_index": false, "create_index": true, "refresh_type": "wait_for", "similarity": "cosine", "timeout": 30, "return_embedding": false, "duplicate_documents": "overwrite", "index_type": "flat", "scroll": "1d", "skip_missing_embeddings": true, "synonym_type": "synonym", "use_system_proxy": false}}, {"name": "tfidf_retriever", "type": "TfidfRetriever", "params": {"top_k": 10}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "tfidf_retriever"}, {"inputs": ["tfidf_retriever"]}]}]}

Generate faq system consisting of pinecone document store and FilterRetriever

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "validate_index_sync": true}}, {"name": "filter_retriever", "type": "FilterRetriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "filter_retriever"}, {"inputs": ["filter_retriever"]}]}]}

Generate Haystack ExtractiveQAPipeline with weaviate document store, table reader and filter retriever

{"version": "1.8.0", "components": [{"name": "weaviate_document_store", "type": "WeaviateDocumentStore", "params": {"port": 0, "timeout_config": [5, 15], "index": "Document", "embedding_dim": 768, "content_field": "content", "name_field": "name", "similarity": "cosine", "index_type": "hnsw", "return_embedding": false, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false}}, {"name": "filter_retriever", "type": "FilterRetriever", "params": {"top_k": 10, "all_terms_must_match": false, "scale_score": true}}, {"name": "table_reader", "type": "TableReader", "params": {"model_name_or_path": "google/tapas-base-finetuned-wtq", "use_gpu": true, "top_k": 10, "top_k_per_candidate": 3, "return_no_answer": false, "max_seq_len": 256, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "filter_retriever"}, {"inputs": ["filter_retriever"], "name": "table_reader"}]}]}

Generate Haystack faq search pipeline using deepset cloud document store and table text retriever

{"version": "1.8.0", "components": [{"name": "deepset_cloud_document_store", "type": "DeepsetCloudDocumentStore", "params": {"workspace": "default", "duplicate_documents": "overwrite", "similarity": "dot_product", "return_embedding": false, "label_index": "default", "embedding_dim": 768}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"]}]}]}

Create Haystack QuestionAnswerGenerationPipeline using farm reader and QuestionGenerator

{"version": "1.8.0", "components": [{"name": "question_generator", "type": "QuestionGenerator", "params": {"sep_token": "<sep>", "batch_size": 16, "progress_bar": true, "use_auth_token": false}}, {"name": "farm_reader", "type": "FARMReader", "params": {"context_window_size": 150, "batch_size": 50, "use_gpu": true, "no_ans_boost": 0.0, "return_no_answer": false, "top_k": 10, "top_k_per_candidate": 3, "top_k_per_sample": 1, "num_processes": 0, "max_seq_len": 256, "doc_stride": 128, "progress_bar": true, "duplicate_filtering": 0, "use_confidence_scores": true, "confidence_threshold": 0, "use_auth_token": false}}], "pipelines": [{"name": "my_question_answering_pipeline", "nodes": [{"inputs": ["Query"], "name": "question_generator"}, {"inputs": ["question_generator"], "name": "farm_reader"}]}]}

Create Haystack search summarization pipeline with transformers summarizer, elasticsearch filter only retriever and in memory document store

{"version": "1.8.0", "components": [{"name": "in_memory_document_store", "type": "InMemoryDocumentStore", "params": {"index": "document", "label_index": "label", "embedding_dim": 768, "return_embedding": false, "similarity": "dot_product", "progress_bar": true, "duplicate_documents": "overwrite", "use_gpu": true, "scoring_batch_size": 500000}}, {"name": "elasticsearch_filter_only_retriever", "type": "ElasticsearchFilterOnlyRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}, {"name": "transformers_summarizer", "type": "TransformersSummarizer", "params": {"model_name_or_path": "google/pegasus-xsum", "max_length": 200, "min_length": 5, "use_gpu": true, "clean_up_tokenization_spaces": true, "separator_for_single_summary": " ", "generate_single_summary": false, "batch_size": 16, "progress_bar": true, "use_auth_token": false}}], "pipelines": [{"name": "my_query_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_filter_only_retriever"}, {"inputs": ["elasticsearch_filter_only_retriever"], "name": "transformers_summarizer"}]}]}

Build search pipeline consisting of pinecone document store and ElasticsearchRetriever

{"version": "1.8.0", "components": [{"name": "pinecone_document_store", "type": "PineconeDocumentStore", "params": {"environment": "us-west1-gcp", "embedding_dim": 768, "return_embedding": false, "index": "document", "similarity": "cosine", "replicas": 1, "shards": 1, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false, "validate_index_sync": true}}, {"name": "elasticsearch_retriever", "type": "ElasticsearchRetriever", "params": {"top_k": 10, "all_terms_must_match": false}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "elasticsearch_retriever"}, {"inputs": ["elasticsearch_retriever"]}]}]}

Generate Haystack faq system with TableTextRetriever and weaviate document store

{"version": "1.8.0", "components": [{"name": "weaviate_document_store", "type": "WeaviateDocumentStore", "params": {"port": 0, "timeout_config": [5, 15], "index": "Document", "embedding_dim": 768, "content_field": "content", "name_field": "name", "similarity": "cosine", "index_type": "hnsw", "return_embedding": false, "embedding_field": "embedding", "progress_bar": true, "duplicate_documents": "overwrite", "recreate_index": false}}, {"name": "table_text_retriever", "type": "TableTextRetriever", "params": {"max_seq_len_query": 64, "max_seq_len_passage": 256, "max_seq_len_table": 256, "top_k": 10, "use_gpu": true, "batch_size": 16, "embed_meta_fields": ["name", "section_title", "caption"], "use_fast_tokenizers": true, "similarity_function": "dot_product", "global_loss_buffer_size": 150000, "progress_bar": true, "use_auth_token": false, "scale_score": true, "use_fast": true}}], "pipelines": [{"name": "my_doc_search_pipeline", "nodes": [{"inputs": ["Query"], "name": "table_text_retriever"}, {"inputs": ["table_text_retriever"]}]}]}