Metadata Filter Extraction With 2 CoT Examples

Your goal is to structure the user's query to match the request schema provided below.

When responding use a markdown code snippet with a JSON object formatted in the following schema:

{schema_str}

The query string should contain only text that is expected to match the contents of documents. Any conditions in the filter should not be mentioned in the query as well.

Restrictions

• Make sure that filters only refer to attributes that exist in the data source.
• Make sure that filters take into account the descriptions of attributes.
• MAKE SURE THAT FILTERS ARE ONLY USED AS NEEDED. IF THERE - ARE NO FILTERS THAT SHOULD BE APPLIED RETURN [] FOR THE FILTER VALUE.
• Do NOT EVER infer a null value for a filter. This will break the downstream program. Instead, don't include the filter.
• If more than one value will be used for one filter do not use multiple == for the same filter key. For example use ...⟨"key": "report_quarter", "value": ["Q1", "Q2"], "operator": "in"⟩.... DO NOT USE ...⟨"key": "report_quarter", "value" Q1, "operator": "=="⟩, ⟨"key": "report_quarter", "value" Q2, "operator": "=="⟩...

Data Source:

⟨
    "metadata_info": [
       {{
           "name"="trading_symbol",
           "description"="Four character trading symbol of the company. AAPL for Apple, AMZN for Amazon, INTC for Intel, MSFT for Microsoft, and NVDA for Nvidia",
           "type"="string",
       ⟩,
       ⟨
            "name"="report_year",
           "description"="Creation year of the document.",
           "type"="string",
       ⟩,
       ⟨
            "name"="report_quarter",
            "description"="Creation quarter of the document. Only can be one of Q1, Q2, or Q3",
           "type"="string",
       ⟩,
    ],
    "content_info": "SEC FORM-10Q Documents"
}}

User Query: In Amazon's Q2 2023 10-Q, how does the discussion on employee compensation and benefits in the human resources section relate to the reported financial expenses?

Additional Instructions: If one of the filters is 'trading_symbol', please make sure that all letters of the inferred value is capitalized. Only words that are capitalized are valid values for "trading_symbol".

Structured Request:

⟨
    "query": "Discussion on employee compensation and benefits in the human resources section relate to the reported financial expenses",
     "filters": [
          {{"key": "trading_symbol", "value": "AMZN", "operator": "=="⟩,
          ⟨"key": "report_year", "value": "2023", "operator": "=="⟩,
          ⟨"key": "report_quarter", "value" Q2, "operator": "=="⟩
     ],
}}

Data Source:

{info_str}

User Query: {query_str}

Additional Instructions: {additional_instructions}

Structured Request: