Add search method to PipelineResource

[mathislucka]

Overview
First check REPO.md to see general contribution guidelines.
We want to add the ability to use a pipeline for search to the PipelineResource. To do that, we will use the search or search_stream endpoint of the deepset platform (when streaming=True).

Context

This is a request to the search endpoint:

curl --request POST \
     --url https://api.cloud.deepset.ai/api/v1/workspaces/some_workspace/pipelines/some_pipe/search \
     --header 'accept: application/json' \
     --header 'content-type: application/json' \
     --data '
{
  "debug": false,
  "view_prompts": false,
  "queries": [
    "myquery"
  ]
}
'

Note that you can send multiple queries in one request but we do NOT want to expose that feature in the PipelineResource. We only ever want to send a single query.

You can also add filters and additional params to the request. The additional params are additional inputs for components within the Pipeline we are searching. Below is an example:

curl --request POST \
     --url https://api.cloud.deepset.ai/api/v1/workspaces/some_workspace/pipelines/some_pipe/search \
     --header 'accept: application/json' \
     --header 'content-type: application/json' \
     --data '
{
  "debug": false,
  "params": {
    "some_component": "{\"some_input\":  123}"
  },
  "view_prompts": false,
  "queries": [
    "myquery"
  ],
  "filters": {
    "conditions": [
      {
        "field": "filename",
        "value": "test.txt"
      },
      {
        "field": "date_created",
        "operator": "<=",
        "value": "22.12.2023"
      }
    ]
  }
}
'

This is an example response:

{
  "query_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
  "results": [
    {
      "_debug": {
        "additionalProp": {}
      },
      "answers": [
        {
          "answer": "string",
          "context": "string",
          "document_id": "string",
          "document_ids": [
            "string"
          ],
          "file": {
            "additionalProp": {}
          },
          "files": [
            {
              "additionalProp": {}
            }
          ],
          "meta": {
            "additionalProp": {}
          },
          "offsets_in_context": [
            {
              "end": 0,
              "start": 0
            }
          ],
          "offsets_in_document": [
            {
              "end": 0,
              "start": 0
            }
          ],
          "prompt": "string",
          "result_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
          "score": 0,
          "type": "generative"
        }
      ],
      "documents": [
        {
          "content": "string",
          "embedding": [
            0
          ],
          "file": {
            "additionalProp": {}
          },
          "id": "string",
          "meta": {
            "additionalProp": {}
          },
          "result_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
          "score": 0
        }
      ],
      "prompts": {
        "additionalProp": "string"
      },
      "query": "string",
      "query_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6"
    }
  ]
}

The general structure of the result is that we have answers and documents. For answers, the only required field is answer. Additional fields may vary.

For documents, we always have content and meta. Meta is a dict that can hold any value. score and embedding are optional.

Prompts will expose one or more prompts used in the pipeline. It is a mapping of component_name: prompt. Prompts is optional.

_debug holds debug state of the pipeline (outputs and inputs of every component). This can be very long.

This is what full debug output might look like:

{
  "bm25_retriever": {
    "documents": [
      {
        "id": "1ccb9ba8a25a6e929a15510db70b1a0ff12f7aa722dd5976170ccda9e0be241d",
        "content": "This led to a $700 million or\n\n\n50 basis point headwind to revenue relative to what we guided. Excluding this FX\n\n\nheadwind, we would have exceeded the top end of our guidance range.\n\n\nWorldwide operating income was $15.3 billion, which was our highest quarterly\n\n\nincome ever, and it was $3.3 billion above the high end of our guidance range. This\n\n\nwas driven by strong operational performance across all three reportable segments\n\n\nand better-than-expected operating leverage, including lower cost to serve. The\n\n\nimpact on operating income from our Q1 FX rate headwind was negligible. I'll speak\n\n\nmore to our pro\u0000tability trends in a moment.\n\n\nIn the North America segment, \u0000rst quarter revenue was $86.3 billion, an increase of\n\n\n12% year over year. In the international segment, revenue was $31.9 billion, an\n\n\nincrease of 11% year over year, excluding the impact of foreign exchange. We\n\n\nremain focused on the inputs that matter most to our customers: selection, price,\n\n\nand convenience. During the quarter, around the world, we help customers save with\n\n\nour shopping events.\n\n\nWe added selection, including premium and luxury brands, and we delivered our\n\n\nfastest speeds ever for Prime members. Third-party sellers continue to be an\n\n\nimportant part of our offering. Third-party seller services revenue increased 16%\n\n\nyear over year, excluding the impact of foreign exchange. ",
        "blob": null,
        "meta": {
          "_split_overlap": [
            {
              "range": [
                1364,
                1549
              ],
              "doc_id": "7b9e0694b7fa5f8eb39405c07c47cc19057c667a2f3ec96e28194d30f7f04267"
            },
            {
              "range": [
                0,
                301
              ],
              "doc_id": "04a662e2161b55fdc6a432dd502c3170b42a38f699863e0ef320221c96a26304"
            }
          ],
          "split_idx_start": 21629,
          "file_name": "Amazon.com (AMZN) Q1 2024 Earnings Call Transcript _ The Motley Fool.pdf",
          "_file_created_at": "2024-06-02T11:26:35.503979+00:00",
          "split_id": 17,
          "_file_size": 255003,
          "page_number": 10,
          "file_id": "672d6263-03c4-4f87-a09d-b8b90ad6d3f5",
          "source_id": "b63bc5e45d744892944acf5e6c2aeee189ec4eef434848f94933e317a20dc06b"
        },
        "score": 5.7458057,
        "embedding": null,
        "sparse_embedding": null
      }
]
  },
  "embedding_retriever": {
    "documents": [
      {
        "id": "9715ff6f92656a6145273e34d21810d5511e1b30232ff15685c6eaba46e6eec9",
        "content": "Net income was $12.4 billion or $4.71 per share. Capital expenditures\n\n\nincluding principal payments on finance leases were $6.7 billion driven by investments in servers, data centers and network\n\n\ninfrastructure. Free cash flow was $12.5 billion. We repurchased $14.6 billion of our Class A common stock and paid $1.3 billion\n\n\nin dividends to shareholders ending the quarter with $58.1 billion in cash and marketable securities and $18.4 billion in debt.\n\n\nMoving now to our segment results.\n\n\nI'll begin with our Family of Apps segment. Our community across the Family of Apps continues to grow with approximately 3.2\n\n\nbillion people using at least one of our Family of Apps on a daily basis in March. Q1 total Family of Apps revenue was $36 billion,\n\n\nup 27% year-over-year. Q1 Family of Apps ad revenue was $35.6 billion, up 27% or 26% on a constant currency basis. Within ad\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n4/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\nrevenue, the online commerce vertical was the largest contributor to year-over-year growth followed by gaming and\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nENTERTAINMENT\n\n\nSHOPPING\n\n\nSEARCH\n\n\n HOME\n\n\nLIFE\n\n\nentertainment and media. On a user geography basis, ad revenue growth was strongest in Rest of World and Europe at 40%\n\n\nand 33%, respectively. Asia-Pacific grew 25% and North America grew 22%. In Q1, the total number of ad impressions served\n\n\nSign in\n\n\nMail\n\n\nacross our services increased 20% and the average price per ad increased 6%.\n\n\n",
        "blob": null,
        "meta": {
          "_split_overlap": [
            {
              "range": [
                1251,
                1497
              ],
              "doc_id": "a535d3ef57134d9174c14b8cb3e55ea4c5ccf278ac2f6c0e7eef801f98b0284f"
            },
            {
              "range": [
                0,
                196
              ],
              "doc_id": "8f248f0605d812bc7f991820fdb0799070218fd4fd2e17569c597a1c460bd2a0"
            }
          ],
          "split_idx_start": 16435,
          "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
          "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
          "split_id": 15,
          "_file_size": 1386987,
          "page_number": 4,
          "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
          "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
        },
        "score": 1.8627398,
        "embedding": null,
        "sparse_embedding": null
      },
    ]
  },
  "document_joiner": {
    "documents": [
      {
        "id": "6d421f8aa4ce5f7a98baca506c806d98e9b5a1ee1ef6cf2ed76d13d222e8c95d",
        "content": "In Office Commercial, revenue growth will again be driven by Office 365 with\n\n\nseat growth across customer segments and ARPU growth thru E5. We expect\n\n\nOffice 365 revenue growth to be roughly 16% in constant currency. We’re\n\n\nexcited for Microsoft 365 Copilot general availability on November 1 and\n\n\nexpect the related revenue to grow gradually over time. In our on-premises\n\n\nbusiness, we expect revenue to decline in the mid to high teens.\n\n\nIn Office consumer, we expect revenue growth in the mid-single digits,\n\n\ndriven by Microsoft 365 subscriptions.\n\f\n\nFor LinkedIn, we expect revenue growth in the mid-single digits driven by\n\n\nTalent Solutions and Marketing Solutions. Growth continues to be impacted\n\n\nby the overall market environments for recruiting and advertising, especially\n\n\nin the technology industry where we have significant exposure.\n\n\nAnd in Dynamics, we expect revenue growth in the high teens driven by\n\n\nDynamics 365.\n\n\nFor Intelligent Cloud we expect revenue to grow between 17 and 18%, or $25.1\n\n\nto $25.4 billion US dollars. Revenue growth in constant currency will be\n\n\napproximately one point lower.\n\n\nRevenue will continue to be driven by Azure which, as a reminder, can have\n\n\nquarterly variability primarily from our per-user business and from in-period\n\n\nrevenue recognition depending on the mix of contracts.\n\n\nIn Azure, we expect revenue growth to be 26% to 27% in constant currency\n\n\nwith increasing contribution from AI. Growth continues to be driven by Azure\n\n\nconsumption business and we expect the trends from Q1 to continue into Q2.\n\n\n",
        "blob": null,
        "meta": {
          "_split_overlap": [
            {
              "range": [
                1278,
                1494
              ],
              "doc_id": "e6859ad3da078c91daebaef824df4589307249511834aced345bc71dc4f15776"
            },
            {
              "range": [
                0,
                231
              ],
              "doc_id": "29684ca99da5a2c3dc72e16ab63ffa977450f3d57e5543df280fc95309b271a5"
            }
          ],
          "split_idx_start": 28721,
          "file_name": "Microsoft_TranscriptFY24Q1.docx.pdf",
          "_file_created_at": "2024-06-02T11:32:22.294877+00:00",
          "split_id": 22,
          "_file_size": 195483,
          "page_number": 20,
          "file_id": "d26e1af7-10e7-481a-8d74-90e358545ecf",
          "source_id": "7c0ab6310da68f731cc4333c28cc7335ad366f5197f9ec07b0f160a10e7fa841"
        },
        "score": 1.813297,
        "embedding": null,
        "sparse_embedding": null
    }
    ]
  },
  "ranker": {
    "documents": [
      {
        "id": "9715ff6f92656a6145273e34d21810d5511e1b30232ff15685c6eaba46e6eec9",
        "content": "Net income was $12.4 billion or $4.71 per share. Capital expenditures\n\n\nincluding principal payments on finance leases were $6.7 billion driven by investments in servers, data centers and network\n\n\ninfrastructure. Free cash flow was $12.5 billion. We repurchased $14.6 billion of our Class A common stock and paid $1.3 billion\n\n\nin dividends to shareholders ending the quarter with $58.1 billion in cash and marketable securities and $18.4 billion in debt.\n\n\nMoving now to our segment results.\n\n\nI'll begin with our Family of Apps segment. Our community across the Family of Apps continues to grow with approximately 3.2\n\n\nbillion people using at least one of our Family of Apps on a daily basis in March. Q1 total Family of Apps revenue was $36 billion,\n\n\nup 27% year-over-year. Q1 Family of Apps ad revenue was $35.6 billion, up 27% or 26% on a constant currency basis. Within ad\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n4/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\nrevenue, the online commerce vertical was the largest contributor to year-over-year growth followed by gaming and\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nENTERTAINMENT\n\n\nSHOPPING\n\n\nSEARCH\n\n\n HOME\n\n\nLIFE\n\n\nentertainment and media. On a user geography basis, ad revenue growth was strongest in Rest of World and Europe at 40%\n\n\nand 33%, respectively. Asia-Pacific grew 25% and North America grew 22%. In Q1, the total number of ad impressions served\n\n\nSign in\n\n\nMail\n\n\nacross our services increased 20% and the average price per ad increased 6%.\n\n\n",
        "blob": null,
        "meta": {
          "_split_overlap": [
            {
              "range": [
                1251,
                1497
              ],
              "doc_id": "a535d3ef57134d9174c14b8cb3e55ea4c5ccf278ac2f6c0e7eef801f98b0284f"
            },
            {
              "range": [
                0,
                196
              ],
              "doc_id": "8f248f0605d812bc7f991820fdb0799070218fd4fd2e17569c597a1c460bd2a0"
            }
          ],
          "split_idx_start": 16435,
          "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
          "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
          "split_id": 15,
          "_file_size": 1386987,
          "page_number": 4,
          "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
          "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
        },
        "score": 0.46159314047502403,
        "embedding": null,
        "sparse_embedding": null
      }
    ]
  },
  "meta_field_grouping_ranker": {
    "documents": [
      {
        "id": "6816a46ffb464714df712c96d21842f82a1eed59700abccf1445be3ada3e3bd1",
        "content": "\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\n HOME\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nENTERTAINMENT\n\n\nLIFE\n\n\nSEARCH\n\n\nSHOPPING\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nSearch for news, symbols or companies\n\n\nSign in\n\n\nMail\n\n\nS&P 500\n\n\n5,277.51\n+42.03 (+0.80%)\n\n\nDow 30\n\n\n38,686.32\n+574.84 (+1.51%)\n\n\nNasdaq\n\n\n16,735.02\n-2.06 (-0.01%)\n\n\nRussell 2000\n\n\n2,070.13\n+13.53 (+0.66%)\n\n\nCrude Oil\n\n\n77.18\n-0.73 (-0.94%)\n\n\nU.S. markets closed\n\n\nGold\n\n\n2,347.70\n-18.80 (-0\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings\nCall Transcript\n\n\nInsider Monkey Transcripts\nFri, Apr 26, 2024 • 21 min read\n\n\nIn this article:\n\n\nMETA -0.05%\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript April 24, 2024\n\n\nMeta Platforms, Inc. beats earnings expectations. Reported EPS is $4.71, expectations were $4.32.\n\n\nOperator: Good afternoon. My name is Krista and I will be your conference operator today. At this time, I would like to\n\n\nwelcome everyone to the Meta First Quarter Earnings Conference Call. All lines have been placed on mute to prevent any\n\n\nbackground noise. After the speakers' remarks there will be a question-and-answer session. [Operator Instructions] This call will\n\n\nbe recorded. Thank you very much. Ken Dorell, Meta's Director of Investor Relations, you may begin.\n\n\nKenneth Dorell: Thank you. Good afternoon and welcome to Meta Platform's first quarter 2024 earnings conference call.\n\n\nJoining me today to discuss our results are Mark Zuckerberg, CEO; and Susan Li, CFO. Before we get started, I would like to take\n\n\nthis opportunity to remind you that our remarks today will include forward-looking statements. ",
        "blob": null,
        "meta": {
          "_split_overlap": [
            {
              "range": [
                0,
                226
              ],
              "doc_id": "41bb91e0bc75f53f26cadc985895e573ba122b4f36ca0c7c7e4464a7551c87b0"
            }
          ],
          "split_idx_start": 0,
          "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
          "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
          "split_id": 0,
          "_file_size": 1386987,
          "page_number": 1,
          "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
          "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
        },
        "score": 0.0201002355767521,
        "embedding": null,
        "sparse_embedding": null
      }
    ]
  },
  "prompt_builder": {
    "prompt": "You are a technical expert.\nYou answer questions truthfully based on provided documents.\nIf the answer exists in several documents, summarize them.\nIgnore documents that don't contain the answer to the question.\nOnly answer based on the documents provided. Don't make things up.\nIf no information related to the question can be found in the document, say so.\nAlways use references in the form [NUMBER OF DOCUMENT] when using information from a document, e.g. [3] for Document [3] .\nNever name the documents, only enter a number in square brackets as a reference.\nThe reference must only refer to the number that comes in square brackets after the document.\nOtherwise, do not use brackets in your answer and reference ONLY the number of the document without mentioning the word document.\n\nThese are the documents:\nDocument [1] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\n\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\n HOME\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nENTERTAINMENT\n\n\nLIFE\n\n\nSEARCH\n\n\nSHOPPING\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nSearch for news, symbols or companies\n\n\nSign in\n\n\nMail\n\n\nS&P 500\n\n\n5,277.51\n+42.03 (+0.80%)\n\n\nDow 30\n\n\n38,686.32\n+574.84 (+1.51%)\n\n\nNasdaq\n\n\n16,735.02\n-2.06 (-0.01%)\n\n\nRussell 2000\n\n\n2,070.13\n+13.53 (+0.66%)\n\n\nCrude Oil\n\n\n77.18\n-0.73 (-0.94%)\n\n\nU.S. markets closed\n\n\nGold\n\n\n2,347.70\n-18.80 (-0\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings\nCall Transcript\n\n\nInsider Monkey Transcripts\nFri, Apr 26, 2024 • 21 min read\n\n\nIn this article:\n\n\nMETA -0.05%\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript April 24, 2024\n\n\nMeta Platforms, Inc. beats earnings expectations. Reported EPS is $4.71, expectations were $4.32.\n\n\nOperator: Good afternoon. My name is Krista and I will be your conference operator today. At this time, I would like to\n\n\nwelcome everyone to the Meta First Quarter Earnings Conference Call. All lines have been placed on mute to prevent any\n\n\nbackground noise. After the speakers' remarks there will be a question-and-answer session. [Operator Instructions] This call will\n\n\nbe recorded. Thank you very much. Ken Dorell, Meta's Director of Investor Relations, you may begin.\n\n\nKenneth Dorell: Thank you. Good afternoon and welcome to Meta Platform's first quarter 2024 earnings conference call.\n\n\nJoining me today to discuss our results are Mark Zuckerberg, CEO; and Susan Li, CFO. Before we get started, I would like to take\n\n\nthis opportunity to remind you that our remarks today will include forward-looking statements. \n\nDocument [2] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nIt's been a good start to the year both in terms of\n\n\nproduct momentum and business performance. We estimate that more than 3.2 billion people use at least one of our apps\n\n\neach day and we're seeing healthy growth in the US. And I want to call out WhatsApp specifically where the number of daily\n\n\nactives and message sends in the US keeps gaining momentum and I think we're on a good path there. We've also made good\n\n\nprogress on our AI and Metaverse efforts and that's where I'm going to focus most of my comments today. So let's start with AI.\n\n\nWe are building a number of different AI services from Meta AI, our AI assistant that you can ask any question across our apps\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n1/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\nand glasses, to creator AIs that help creators engage their communities and that fans can interact with, to business AIs that we\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nENTERTAINMENT\n\n\nSHOPPING\n\n\nSEARCH\n\n\n HOME\n\n\nLIFE\n\n\nthink every business eventually on our platform will use to help customers buy things and get customer support, to internal\n\n\ncoding and development AIs, to hardware like glasses for people to interact with AIs and a lot more.\n\n\n\n\nDocument [3] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nWe are building a number of different AI services from Meta AI, our AI assistant that you can ask any question across our apps\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n1/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\nand glasses, to creator AIs that help creators engage their communities and that fans can interact with, to business AIs that we\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nENTERTAINMENT\n\n\nSHOPPING\n\n\nSEARCH\n\n\n HOME\n\n\nLIFE\n\n\nthink every business eventually on our platform will use to help customers buy things and get customer support, to internal\n\n\ncoding and development AIs, to hardware like glasses for people to interact with AIs and a lot more.\n\n\nSign in\n\n\nMail\n\n\nLast week, we had the major release of our new version of Meta AI that is now powered by our latest model, Llama 3 and our\n\n\ngoal with Meta AI is to build the world's leading AI service, both in quality and usage. The initial rollout of Meta AI is going well.\n\n\nTens of millions of people have already tried it. The feedback is very positive and when I first checked in with our teams, the\n\n\nmajority of feedback we were getting was people asking us to release Meta AI for them wherever they are. So we've started\n\n\nlaunching Meta AI in some English speaking countries and we'll roll out in more languages and countries over the coming\n\n\nmonths. You all know our product development playbook by this point. \n\nDocument [4] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nThere are several\n\n\nways to build a massive business here, including scaling business messaging, introducing ads or paid content into AI\n\n\ninteractions and enabling people to pay to use bigger AI models and access more compute. And on top of those. AI is already\n\n\nhelping us improve app engagement, which naturally leads to seeing more ads and improving ads directly to deliver more value.\n\n\nSo if the technology and products evolve in the way that we hope, each of those will unlock massive amounts of value for\n\n\npeople and business for us over time. We are seeing good progress on some of these efforts already. Right now, about 30% of\n\n\nthe posts on Facebook feed are delivered by our AI recommendation system. That's up 2x over the last couple of years and for\n\n\nthe first time ever, more than 50% of the content that people see on Instagram is now AI recommended. AI has also been a\n\n\nhuge part of how we create value for advertisers by showing people more relevant ads. And if you look at our two end-to-end\n\n\nAI powered tools, Advantage+ shopping and Advantage+ app campaigns, revenue flowing through those has more than\n\n\ndoubled since last year.\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n2/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\nWe're also going to continue to be very focused on efficiency as we scale Meta AI and other AI services. \n\nDocument [5] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nNet income was $12.4 billion or $4.71 per share. Capital expenditures\n\n\nincluding principal payments on finance leases were $6.7 billion driven by investments in servers, data centers and network\n\n\ninfrastructure. Free cash flow was $12.5 billion. We repurchased $14.6 billion of our Class A common stock and paid $1.3 billion\n\n\nin dividends to shareholders ending the quarter with $58.1 billion in cash and marketable securities and $18.4 billion in debt.\n\n\nMoving now to our segment results.\n\n\nI'll begin with our Family of Apps segment. Our community across the Family of Apps continues to grow with approximately 3.2\n\n\nbillion people using at least one of our Family of Apps on a daily basis in March. Q1 total Family of Apps revenue was $36 billion,\n\n\nup 27% year-over-year. Q1 Family of Apps ad revenue was $35.6 billion, up 27% or 26% on a constant currency basis. Within ad\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n4/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\nrevenue, the online commerce vertical was the largest contributor to year-over-year growth followed by gaming and\n\n\nMAIL\n\n\nNEWS\n\n\nFINANCE\n\n\nSPORTS\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nENTERTAINMENT\n\n\nSHOPPING\n\n\nSEARCH\n\n\n HOME\n\n\nLIFE\n\n\nentertainment and media. On a user geography basis, ad revenue growth was strongest in Rest of World and Europe at 40%\n\n\nand 33%, respectively. Asia-Pacific grew 25% and North America grew 22%. In Q1, the total number of ad impressions served\n\n\nSign in\n\n\nMail\n\n\nacross our services increased 20% and the average price per ad increased 6%.\n\n\n\n\nDocument [6] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nIn April, we rolled out this unified video experience in the US and Canada, which is increasingly\n\n\npowered by our next generation ranking architecture that we expect will help deliver more relevant video recommendations\n\n\nover time. We're also introducing deeper integrations of Generative AI into our apps in the US and more than a dozen other\n\n\ncountries.\n\n\nAlong with using Meta AI within our chat surfaces, people will now be able to use Meta AI in search within our apps as well as\n\n\nfeed and groups on Facebook. We expect these integrations will complement our social discovery strategy as our\n\n\nrecommendation systems help people to discover and explore their interests while Meta AI enables them to dive deeper on\n\n\ntopics they're interested in. Threads also continues to see good traction as we continue to ship valuable features and scale the\n\n\ncommunity. Now to the second driver of our revenue performance, increasing monetization efficiency. There are two parts to\n\n\nthis work. The first is optimizing the level of ads within organic engagement. Here we continue to advance our understanding of\n\n\nusers' preferences for viewing ads to more effectively optimize the right time, place and person to show an ad to.\n\n\nFor example, we are getting better at adjusting the placement and number of ads in real time based on our perception of a\n\n\nuser's interest in ad content and to minimize disruption from ads as well as innovating on new and creative ad formats. \n\nDocument [7] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nWe are also starting to see our AI\n\n\nSign in\n\n\nMail\n\n\ninitiatives increasingly overlap with our Reality Labs work.\n\n\nFor example, with Ray-Ban Meta smart glasses, people in the US and Canada can now use our multimodal meta AI assistant for\n\n\ndaily tasks without pulling out their phone. Longer term, we expect Generative AI to play an increasing role in our mixed reality\n\n\nproducts making it easier to develop immersive experiences. Accelerating our AI efforts will help ensure we can provide the\n\n\nbest version of our services as we transition to the next computing platform. We expect to pursue these opportunities while\n\n\nmaintaining a focus on operating discipline and we believe our strong financial position will allow us to support these\n\n\ninvestments while also returning capital to shareholders through share repurchases and dividends. In addition, we continue to\n\n\nmonitor an active regulatory landscape, including the increasing legal and regulatory headwinds in the EU and the US that\n\n\ncould significantly impact our business and our financial results.\n\n\nWe also have a jury trial scheduled for June in a suit brought by the State of Texas regarding our use of facial recognition\n\n\ntechnology, which could ultimately result in a material loss. Turning now to the revenue outlook. We expect second quarter\n\n\n2024 total revenue to be in the range of $36.5 billion to $39 billon. Our guidance assumes foreign currency as a 1% headwind to\n\n\nyear-over-year total revenue growth based on current exchange rates. Turning now to the expense outlook. \n\nDocument [8] :\nName of Source File: Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf\nOn to tax. Absent any changes to our tax landscape, we expect our full year 2024 tax rate to be in the mid-teens. In\n\n\nclosing, Q1 was a good start to the year. We're seeing strong momentum within our Family of Apps and are making important\n\n\nprogress on our longer term AI and Reality Labs initiatives that have the potential to transform the way people interact with\n\n\nour services over the coming years.\n\n\nWith that, Krista, let's open up the call for questions.\n\n\nSee also\n\n\n20 Countries with the Highest Consumer Spending in the World and\n\n\n12 Countries with the Largest Slum Population in Europe.\n\n\nTo continue reading the Q&A session, please click here.\n\n\n View comments\n\n\nTerms/Privacy Policy\n\n\nPrivacy & Cookie Settings\n\n\nAbout Our Ads\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n6/7\n\f\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\n HOME\n\n\nMAIL\n\n\nNEWS\nFeatured Broker   \n\n\nFINANCE\n\n\nSPORTS\n\n\nENTERTAINMENT\n\n\nLIFE\n\n\nSEARCH\n\n\nSHOPPING\n\n\nYAHOO PLUS\n\n\nMORE...\n\n\nTrust TrustPilot\nTrade with a 4.4/5 TrustPilot rated broker. 20+K users worldwide rated eToro as\nexcellent.\n\n\nJoin Now\n\n\nAd eToro\n\n\nSign in\n\n\nMail\n\n\nCopy Trading does not amount to investment advice. The value of your investments may go up or down. Your capital is at risk.\n\n\nComments\n\n\nCommenting on this article has ended\n\n\nLog in Sign up\n\n\nPowered by\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n7/7\n\n\nQuestion: what was meta revenue\nAnswer:",
    "runtime": {
      "prompts_used": [
        "You are a technical expert.\nYou answer questions …meta revenue\nAnswer:"
      ]
    }
  },
  "llm": {
    "replies": [
      "Meta's total revenue for the first quarter was $36 billion [5]."
    ],
    "meta": [
      {
        "model": "gpt-4-turbo-2024-04-09",
        "index": 0,
        "finish_reason": "stop",
        "completion_start_time": "2025-06-12T09:14:51.838012",
        "usage": null
      }
    ]
  },
  "answer_builder": {
    "answers": [
      {
        "data": "Meta's total revenue for the first quarter was $36 billion .",
        "query": "what was meta revenue",
        "documents": [
          {
            "id": "6816a46ffb464714df712c96d21842f82a1eed59700abccf1445be3ada3e3bd1",
            "content": "\n\n02/06/2024, 12:08\n\n\nMeta Platforms, Inc. (NASDAQ:META) Q1 2024 Earnings Call Transcript\n\n\n HOME\n\n\nMAIL\n\n\nNE…d welcome to Meta Platform's first quarter 2024 earnings conference call.\n\n\nJoining me today to discuss our results are Mark Zuckerberg, CEO; and Susan Li, CFO. Before we get started, I would like to take\n\n\nthis opportunity to remind you that our remarks today will include forward-looking statements. ",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    0,
                    226
                  ],
                  "doc_id": "41bb91e0bc75f53f26cadc985895e573ba122b4f36ca0c7c7e4464a7551c87b0"
                }
              ],
              "split_idx_start": 0,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 0,
              "_file_size": 1386987,
              "page_number": 1,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.0201002355767521,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "4e215446c06cd8c47eea91b0ed7ed6277920d7e9ae936e335e63d7068b5c2876",
            "content": "It's been… will use to help customers buy things and get customer support, to internal\n\n\ncoding and development AIs, to hardware like glasses for people to interact with AIs and a lot more.\n\n\n",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    1123,
                    1349
                  ],
                  "doc_id": "41bb91e0bc75f53f26cadc985895e573ba122b4f36ca0c7c7e4464a7551c87b0"
                },
                {
                  "range": [
                    0,
                    770
                  ],
                  "doc_id": "385af76adce965ee05f63c0fd0020d7b598b0af0b7f50644d72362cf97f82f55"
                }
              ],
              "split_idx_start": 2563,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 2,
              "_file_size": 1386987,
              "page_number": 1,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.006413163741325347,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "385af76adce965ee05f63c0fd0020d7b598b0af0b7f50644d72362cf97f82f55",
            "content": "We are b…ver they are. So we've started\n\n\nlaunching Meta AI in some English speaking countries and we'll roll out in more languages and countries over the coming\n\n\nmonths. You all know our product development playbook by this point. ",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    549,
                    1319
                  ],
                  "doc_id": "4e215446c06cd8c47eea91b0ed7ed6277920d7e9ae936e335e63d7068b5c2876"
                },
                {
                  "range": [
                    0,
                    210
                  ],
                  "doc_id": "f705f63465e86542dc0e2bd435102ebc3f6cb658a9c213971ff123dabdd9f8a7"
                }
              ],
              "split_idx_start": 3112,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 3,
              "_file_size": 1386987,
              "page_number": 1,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.0024534282596974846,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "20e9b57e8c0e49626076ef7e20a044a8519cb2a9245d57044fc3bc7e6bb5dfb8",
            "content": "There are… very focused on efficiency as we scale Meta AI and other AI services. ",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    955,
                    1348
                  ],
                  "doc_id": "d8afe880c795fbb531f042a9c9dc6973055ef9a9e1fdbc9840efe7aa5ec81e21"
                },
                {
                  "range": [
                    0,
                    278
                  ],
                  "doc_id": "bc27dc9ae0ad5aa9cbdcf546a77b8ea2d76ddd41d86bf5c46dcb0b3422dbce08"
                }
              ],
              "split_idx_start": 7724,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 7,
              "_file_size": 1386987,
              "page_number": 2,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.013795364421185227,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "9715ff6f92656a6145273e34d21810d5511e1b30232ff15685c6eaba46e6eec9",
            "content": "Net income w… grew 22%. In Q1, the total number of ad impressions served\n\n\nSign in\n\n\nMail\n\n\nacross our services increased 20% and the average price per ad increased 6%.\n\n\n",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    1251,
                    1497
                  ],
                  "doc_id": "a535d3ef57134d9174c14b8cb3e55ea4c5ccf278ac2f6c0e7eef801f98b0284f"
                },
                {
                  "range": [
                    0,
                    196
                  ],
                  "doc_id": "8f248f0605d812bc7f991820fdb0799070218fd4fd2e17569c597a1c460bd2a0"
                }
              ],
              "split_idx_start": 16435,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 15,
              "_file_size": 1386987,
              "page_number": 4,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.46159314047502403,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "00c7925901c7497f3520c3175ab4e518070f08ae69c458b09a3a00fbf8eb8ce3",
            "content": "In April, we ro…e to advance our understanding of\n\n\nusers' preferences for viewing ads to more effectively optimize the right time, place and person to show an ad to.\n\n\nFor example, we are getting better at adjusting the placement and number of ads in real time based on our perception of a\n\n\nuser's interest in ad content and to minimize disruption from ads as well as innovating on new and creative ad formats. ",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    1109,
                    1469
                  ],
                  "doc_id": "e544e50ea0bc9e75b452ce1c688feaa24ada2424e400c1e8a984d7899562d242"
                },
                {
                  "range": [
                    0,
                    244
                  ],
                  "doc_id": "03f349aefe37e1853e5fc3f6b989d556e2084e3751a23a60814c0174585e3e9c"
                }
              ],
              "split_idx_start": 20349,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 18,
              "_file_size": 1386987,
              "page_number": 5,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.006488269596656705,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "5a267df275c4e5a9e520766613f77f5a02412bc44f78ab8d732551a7931c10d5",
            "content": "We are als…logy, which could ultimately result in a material loss. Turning now to the revenue outlook. We expect second quarter\n\n\n2024 total revenue to be in the range of $36.5 billion to $39 billon. Our guidance assumes foreign currency as a 1% headwind to\n\n\nyear-over-year total revenue growth based on current exchange rates. Turning now to the expense outlook. ",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    1456,
                    1743
                  ],
                  "doc_id": "eba88cbb33b0389c4f5ebefcf077921088d48cba80b358f85672f528b8e5140e"
                },
                {
                  "range": [
                    0,
                    262
                  ],
                  "doc_id": "9e3a94fbce36dc487b38eb54e5382b4f34cb3b04022d0bf3248d02ceef6124fb"
                }
              ],
              "split_idx_start": 24313,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 21,
              "_file_size": 1386987,
              "page_number": 6,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.06841291035220819,
            "embedding": null,
            "sparse_embedding": null
          },
          {
            "id": "ae53d349397e3e4bd24a809818994f028c665e1f24af08bab8a84902d491c508",
            "content": "On to tax. …as\nexcellent.\n\n\nJoin Now\n\n\nAd eToro\n\n\nSign in\n\n\nMail\n\n\nCopy Trading does not amount to investment advice. The value of your investments may go up or down. Your capital is at risk.\n\n\nComments\n\n\nCommenting on this article has ended\n\n\nLog in Sign up\n\n\nPowered by\n\n\nhttps://finance.yahoo.com/news/meta-platforms-inc-nasdaq-meta-130843596.html\n\n\n7/7",
            "blob": null,
            "meta": {
              "_split_overlap": [
                {
                  "range": [
                    1164,
                    1325
                  ],
                  "doc_id": "9e3a94fbce36dc487b38eb54e5382b4f34cb3b04022d0bf3248d02ceef6124fb"
                }
              ],
              "split_idx_start": 26767,
              "file_name": "Meta Platforms, Inc. (NASDAQ_META) Q1 2024 Earnings Call Transcript.pdf",
              "_file_created_at": "2024-06-02T11:32:23.876448+00:00",
              "split_id": 23,
              "_file_size": 1386987,
              "page_number": 6,
              "file_id": "dfb7e279-5de5-4f0e-b277-e29e91b521f1",
              "source_id": "871e7dd48ae4f0bc067ec41ab082d8c67097d851e70bc9b49f1fac1c982c8bb6"
            },
            "score": 0.0044510412268872835,
            "embedding": null,
            "sparse_embedding": null
          }
        ],
        "meta": {
          "prompt": "You are a technical expert.\n…meta revenue\nAnswer:",
          "_references": [
            {
              "answer_end_idx": 59,
              "answer_start_idx": 59,
              "document_id": "9715ff6f92656a6145273e34d21810d5511e1b30232ff15685c6eaba46e6eec9",
              "document_position": 5,
              "label": "grounded",
              "score": 0,
              "doc_end_idx": 1620,
              "doc_start_idx": 0,
              "origin": "llm"
            }
          ]
        }
      }
    ]
  }
}

For streaming, this is what a request might look like:

curl --request POST \
     --url https://api.cloud.deepset.ai/api/v1/workspaces/some_workspace/pipelines/some_pipeline/search-stream \
     --header 'accept: application/json' \
     --header 'content-type: application/json' \
     --data '
{
  "debug": false,
  "include_result": true,
  "view_prompts": false,
  "query": "hello"
}

Include results determines if the full result is returned as the last delta.

Here is an example of a delta event sent:

data: {"query_id": "9ba44483-6e36-49ba-b1ab-f164adba6e7c", "delta": {"text": "The", "meta": {"model": "gpt-4-turbo-2024-04-09", "index": 0, "finish_reason": null, "received_at": "2025-06-12T09:19:02.504905", "deepset_cloud": {"component": "llm"}}}, "type": "delta"}

Below are some docs on streaming deltas:

Run a search using a pipeline and return the answer as stream.

The full result can be accessed as the last stream message if include_result=True.

Event data format where delta, result and error are mutually exclusive:


{
    "query_id": UUID,
    "type": Literal["delta", "result", "error"],
    "delta": Optional[StreamDelta],
    "result": Optional[DeepsetCloudQueryResponse],
    "error": Optional[str],
}
StreamDelta format:


{
    "text": str,
}

result will have the same structure as the search endpoint's response.

Requirements

• Add a search-method to the PipelineResource
• Create models as needed to handle search responses
• Models need to be flexible, don't overdo it. Mainly, we expect either documents or answers (both can be empty) and want to enforce at least the basics for both (answer.answer, document.content, document.meta)
• search should support filtering, debug, streaming, additional params and all other features shown in the context
• create unit tests for the new method in the same file as the existing unit tests for the PipelineResource; follow the same testing approach (faking of the client)
• create integration tests for the new method in the same file as the existing integration tests for the PipelineResource; keep integration tests to the necessary minimum
• do NOT add any tools based on the new method yet, that will come in another PR

[mathislucka]

Great, I can see that you already implemented the full feature. Good job!

Now think extra hard for the last few changes:

• we do not need a model for filters, filters should just be a dict of str, Any
• we use pipe instead of Optional (str | None instead of Optional[str])
• we use builtin types (dict instead of Dict, list instead of List) for our type hints

There seem to be some test failures:

============================= test session starts ==============================
platform linux -- Python 3.11.12, pytest-8.3.5, pluggy-1.6.0
rootdir: /home/runner/work/deepset-mcp-server/deepset-mcp-server
configfile: pyproject.toml
testpaths: test
plugins: anyio-4.9.0, asyncio-0.26.0
asyncio: mode=Mode.AUTO, asyncio_default_fixture_loop_scope=function, asyncio_default_test_loop_scope=function
collected 205 items / 30 deselected / 175 selected

test/unit/api/custom_components/test_custom_components_resource.py .... [ 2%]
test/unit/api/haystack_service/test_haystack_service_resource.py .F... [ 5%]
test/unit/api/indexes/test_index_resource.py .............. [ 13%]
test/unit/api/pipeline/test_pipeline_resource.py ....................... [ 26%]
.......................F.....F... [ 45%]
test/unit/api/pipeline_template/test_pipeline_template_resource.py ..... [ 48%]
.. [ 49%]
test/unit/api/user/test_user_resource.py ... [ 50%]
test/unit/test_async_deepset_client.py ...... [ 54%]
test/unit/tools/test_component_helper.py ... [ 56%]
test/unit/tools/test_custom_components.py ....... [ 60%]
test/unit/tools/test_formatting_utils.py ....... [ 64%]
test/unit/tools/test_haystack_service.py ......... [ 69%]
test/unit/tools/test_indexes.py ................ [ 78%]
test/unit/tools/test_pipeline.py .......................... [ 93%]
test/unit/tools/test_pipeline_template.py ............ [100%]

=================================== FAILURES ===================================
______________________ test_get_component_schema_success _______________________

mock_client = <test.unit.conftest.BaseFakeClient object at 0x7fcea5619550>
mock_successful_schema_response = None

@pytest.mark.asyncio
async def test_get_component_schema_success(
    mock_client: BaseFakeClient,
    mock_successful_schema_response: None,
) -> None:
    """Test successful component schema retrieval."""
    resource = HaystackServiceResource(client=mock_client)
    result = await resource.get_component_schemas()

    assert result == make_component_schema_response()

  assert mock_client.requests[-1] == {

        "method": "GET",
        "endpoint": "v1/haystack/components",
        "headers": {"accept": "application/json"},
        "data": {"domain": "deepset-cloud"},
    }

E AssertionError: assert {'data': {'do...': 'GET', ...} == {'data': {'do...ethod': 'GET'}
E
E Omitting 4 identical items, use -vv to show
E Left contains 1 more item:
E {'stream': False}
E
E Full diff:
E {
E 'data': {
E 'domain': 'deepset-cloud',
E },
E 'endpoint': 'v1/haystack/components',
E 'headers': {
E 'accept': 'application/json',
E },
E 'method': 'GET',
E + 'stream': False,
E }

test/unit/api/haystack_service/test_haystack_service_resource.py:75: AssertionError
________ TestPipelineSearchResource.test_search_with_debug_and_prompts _________

self = <test.unit.api.pipeline.test_pipeline_resource.TestPipelineSearchResource object at 0x7fcea524f2d0>

@pytest.mark.asyncio
async def test_search_with_debug_and_prompts(self) -> None:
    """Test search with debug and view_prompts enabled."""
    # Create sample search response with debug and prompts
    search_response = {
        "query_id": "123e4567-e89b-12d3-a456-4266[14](https://github.com/deepset-ai/deepset-mcp-server/actions/runs/15607307722/job/43959733337?pr=92#step:6:15)174000",
        "results": [
            {
                "query": "debug query",
                "query_id": "123e4567-e89b-12d3-a456-426614[17](https://github.com/deepset-ai/deepset-mcp-server/actions/runs/15607307722/job/43959733337?pr=92#step:6:18)4000",
                "answers": [],
                "documents": [],
                "prompts": {"llm": "Answer the question: {query}"},
                "_debug": {"llm": {"input": "debug info"}},
            }
        ],
    }

    # Create client with predefined response
    client = DummyClient(responses={"test-workspace/pipelines/test-pipeline/search": search_response})

    # Create resource and call search method with debug and prompts
    resource = PipelineResource(client=client, workspace="test-workspace")
    result = await resource.search(
        pipeline_name="test-pipeline", query="debug query", debug=True, view_prompts=True
    )

    # Verify results
    assert isinstance(result, SearchResponse)
    assert len(result.results) == 1
    assert result.results[0].prompts == {"llm": "Answer the question: {query}"}

  assert result.results[0]._debug == {"llm": {"input": "debug info"}}

E AssertionError: assert None == {'llm': {'input': 'debug info'}}
E + where None = SearchResult(answers=[], documents=[], prompts={'llm': 'Answer the question: {query}'}, query='debug query', query_id=UUID('123e4567-e89b-12d3-a456-426614174000'))._debug

test/unit/api/pipeline/test_pipeline_resource.py:1175: AssertionError
__________ TestPipelineSearchResource.test_search_with_all_parameters __________

self = <test.unit.api.pipeline.test_pipeline_resource.TestPipelineSearchResource object at 0x7fcea5278a10>

@pytest.mark.asyncio
async def test_search_with_all_parameters(self) -> None:
    """Test search with all parameters combined."""
    # Create comprehensive search response
    search_response = {
        "query_id": "123e4567-e89b-12d3-a456-426614174000",
        "results": [
            {
                "query": "comprehensive query",
                "query_id": "123e4567-e89b-12d3-a456-426614174000",
                "answers": [
                    {
                        "answer": "Comprehensive answer",
                        "context": "Some context",
                        "score": 0.98,
                        "meta": {"confidence": "high"},
                    }
                ],
                "documents": [
                    {
                        "content": "Comprehensive content",
                        "meta": {"source": "comprehensive.txt", "author": "test"},
                        "score": 0.95,
                    }
                ],
                "prompts": {"llm": "Comprehensive prompt: {query}"},
                "_debug": {"retriever": {"retrieved_docs": 10}},
            }
        ],
    }

    # Create client with predefined response
    client = DummyClient(responses={"test-workspace/pipelines/test-pipeline/search": search_response})

    # Create resource and call search method with all parameters
    resource = PipelineResource(client=client, workspace="test-workspace")
    params = {"retriever": '{"top_k": 10}'}
    filters = SearchFilters(conditions=[FilterCondition(field="author", value="test")])
    result = await resource.search(
        pipeline_name="test-pipeline",
        query="comprehensive query",
        debug=True,
        view_prompts=True,
        params=params,
        filters=filters,
    )

    # Verify results
    assert isinstance(result, SearchResponse)
    assert len(result.results) == 1
    search_result = result.results[0]
    assert len(search_result.answers) == 1
    assert search_result.answers[0].answer == "Comprehensive answer"
    assert search_result.answers[0].context == "Some context"
    assert len(search_result.documents) == 1
    assert search_result.documents[0].content == "Comprehensive content"
    assert search_result.documents[0].meta == {"source": "comprehensive.txt", "author": "test"}
    assert search_result.prompts == {"llm": "Comprehensive prompt: {query}"}

  assert search_result._debug == {"retriever": {"retrieved_docs": 10}}

E AssertionError: assert None == {'retriever': {'retrieved_docs': 10}}
E + where None = SearchResult(answers=[Answer(answer='Comprehensive answer', context='Some context', document_id=None, document_ids=Non...: 'Comprehensive prompt: {query}'}, query='comprehensive query', query_id=UUID('123e4567-e89b-12d3-a456-426614174000'))._debug

test/unit/api/pipeline/test_pipeline_resource.py:1358: AssertionError
=============================== warnings summary ===============================
.venv/lib/python3.11/site-packages/pydantic/_internal/_config.py:323
/home/runner/work/deepset-mcp-server/deepset-mcp-server/.venv/lib/python3.11/site-packages/pydantic/_internal/_config.py:323: PydanticDeprecatedSince20: Support for class-based config is deprecated, use ConfigDict instead. Deprecated in Pydantic V2.0 to be removed in V3.0. See Pydantic V2 Migration Guide at https://errors.pydantic.dev/2.11/migration/
warnings.warn(DEPRECATION_MESSAGE, DeprecationWarning)

.venv/lib/python3.11/site-packages/pydantic/_internal/_generate_schema.py:293
.venv/lib/python3.11/site-packages/pydantic/_internal/_generate_schema.py:293
/home/runner/work/deepset-mcp-server/deepset-mcp-server/.venv/lib/python3.11/site-packages/pydantic/_internal/_generate_schema.py:293: PydanticDeprecatedSince20: json_encoders is deprecated. See https://docs.pydantic.dev/2.11/concepts/serialization/#custom-serializers for alternatives. Deprecated in Pydantic V2.0 to be removed in V3.0. See Pydantic V2 Migration Guide at https://errors.pydantic.dev/2.11/migration/
warnings.warn(

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
=========================== short test summary info ============================
FAILED test/unit/api/haystack_service/test_haystack_service_resource.py::test_get_component_schema_success - AssertionError: assert {'data': {'do...': 'GET', ...} == {'data': {'do...ethod': 'GET'}

Omitting 4 identical items, use -vv to show
Left contains 1 more item:
{'stream': False}

Full diff:
{
'data': {
'domain': 'deepset-cloud',
},
'endpoint': 'v1/haystack/components',
'headers': {
'accept': 'application/json',
},
'method': 'GET',

• 'stream': False,
  

  }
  FAILED test/unit/api/pipeline/test_pipeline_resource.py::TestPipelineSearchResource::test_search_with_debug_and_prompts - AssertionError: assert None == {'llm': {'input': 'debug info'}}
• where None = SearchResult(answers=[], documents=[], prompts={'llm': 'Answer the question: {query}'}, query='debug query', query_id=UUID('123e4567-e89b-12d3-a456-426614174000'))._debug
  FAILED test/unit/api/pipeline/test_pipeline_resource.py::TestPipelineSearchResource::test_search_with_all_parameters - AssertionError: assert None == {'retriever': {'retrieved_docs': 10}}
• where None = SearchResult(answers=[Answer(answer='Comprehensive answer', context='Some context', document_id=None, document_ids=Non...: 'Comprehensive prompt: {query}'}, query='comprehensive query', query_id=UUID('123e4567-e89b-12d3-a456-426614174000'))._debug
  =========== 3 failed, 172 passed, 30 deselected, 3 warnings in 1.81s ===========

Look at all the changes you made (models, PipelineResource, unit and integration tests, transport and transport protocol, etc.) and implement the last few bits to get this over the line.