The filters for live inventory are in the format of a JSON object. In a
filter object, each key is either a field to filter on or an operator.
Each value is either a filter condition object or, in the case of
certain operators, an array of filter objects.

A filter condition object can contain the following fields:

• equal_to: Exact value for the filter.
• not_equal_to: Not equal to value for the filter.
• greater_than: Greater than value for the filter.
• greater_than_or_equal_to: Greater than or equal to value for the filter.
• less_than: Less than value for the filter.
• less_than_or_equal_to: Less than or equal to value for the filter.
• in: Array of values for the filter.
• not_in: Array of values not to include in the filter.
• is_null: Boolean to filter for null / not-null values.
• includes: String to include in the filter.
• not_includes: String not to include in the filter.
• starts_with: String to start with in the filter.
• ends_with: String to end with in the filter.

The fields that can be filtered on are:

• price
• yield_to_maturity
• yield_to_worst
• isin
• cusip
• issuer
• sector
• description
• type
• issue_date
• maturity_date
• workout_date
• treasury_benchmark
• sp_rating
• lowest_rating_rank
• coupon
• currency
• country_issue
• country_of_risk
• emerging_status
• liquidity_micro_buy
• liquidity_micro_sell
• liquidity_micro_aggregate
• liquidity_retail_buy
• liquidity_retail_sell
• liquidity_retail_aggregate
• liquidity_institutional_buy
• liquidity_institutional_sell
• liquidity_institutional_aggregate
• Any of the other fields in the reference data model.

The operators that can be used are:

• and: Array of filter objects to be combined with an AND operator.
• not: Filter object to be negated.
• or: Array of filter objects to be combined with an OR operator.

For example, a sample filter object might look like this:

{
  "yield_to_worst": {
    "greater_than": 6.5
  },
  "price": {
    "less_than": 100
  },
  "maturity_date": {
    "less_than": "2030-01-01",
    "greater_than": "2026-01-01"
  },
  "sp_rating": {
    "greater_than": "BBB"
  },
  "issuer": {
    "includes": "Apple"
  },
  "liquidity_institutional_buy": {
    "in": [3, 4, 5]
  },
  "or": [
    {
      "description": {
        "starts_with": "app"
      }
    },
    {
      "description": {
        "starts_with": "united"
      }
    }
  ]
}

This must be encoded as a JSON string in the query parameter.

Filtering by trade_size will only return asks and bids that would
completely fill an order of the specified quantity. More specifically,
it will return quotes where the minimum quantity is less than or equal
to the specified trade_size and the total quantity is greater than or
equal to the specified trade_size. Optionally, set value to "lowest" to
get all the quotes sorted from least to greatest min_quantity.

JSON-encoded objective function that defines the scoring formula for ranking bonds.
When provided alongside objective_score in returns, the server computes a score for each bond.
Currently returns a stub value of 0 — the real computation is connected in a follow-up.

Specify a side for inventory filters. Possible values:

• 'buy' will limit the results to bonds that have quotes to buy
  the instrument.
• 'sell' will limit the results to bonds that have quotes to sell
  the instrument.
• 'both' will limit the results to bonds that have quotes on both
  sides of the market, i.e. at least one quote to buy and one to sell.

Important note: If you are building a UI that leverages this API, you should be
especially cognizant that this value represents inventory side, not trade side.

A client who wants to purchase a bond will look for asks/offers where the
counterparty wants to sell. Hence, they should pass 'sell' to the API.

A client who wants to sell a bond will look for bids where the counterparty
wants to buy. Hence, they should pass 'buy' to the API.

The risk group ID to filter for tradable inventory. If left null,
and a single risk group exists for the organization, it will be
used as the default risk group. If left null, and more than one
risk group exists, the request will fail.