Skip to main content

​
What’s new in Liberator 2.4

Liberator 2.4 centers on more precise queries, clearer dataset health, and safer limits. Headline capabilities include wildcard / pattern filters on the symbol (key) parameter, a Super-Admin-only usage_log dataset of per-query records, and descriptive entitlement-limit errors. The release also adds a per-dataset cap on how many days a single query can span, Concordance-backed dataset catalog cards, the CloudQuant Excel plugin in the Liberator UI, and a round of connection and time-range reliability fixes.

​
Highlights

Symbol and key pattern filters

Use % wildcards on symbols / keys to pull a product family or naming pattern without listing every contract.

Clear entitlement errors

Limit violations say which cap was hit and what to change, instead of a generic “not entitled.”

Query day-range cap

Admins can set a maximum number of days a single query may span on a dataset, so one wide request cannot stall everyone else.

Dataset health cards

Catalog cards show date range, symbol count, status, and last updated — in the UI and via liberator.datasets(details=True).

Excel plugin in Liberator

Download and access the CloudQuant Excel plugin from the Liberator client, without a separately shared installer.

Combined user creation

Create and enroll a user in one flow. Invite emails include the username to sign in with.

Per-query usage log

Super Admins query usage_log like any other dataset — who queried what, which window, rows returned, and outcome.

​
Symbol and key pattern filters

symbols (and the keys synonym) accept SQL LIKE patterns. An unescaped % is a wildcard for any sequence of characters:
You can mix exact keys and patterns in one list. To match a literal percent character in a key, escape it as \%. See Symbols and key fields and the query parameters reference.

​
Clearer entitlement errors and query day-range limits

When a query is rejected because of an entitlement or usage limit, Liberator now returns a specific message: which limit applied, the configured value, and what to change. A row-limit miss no longer looks the same as a missing entitlement. Super Admins can also set a maximum number of days a single query may span on a given dataset (or dataset group). Use this on high-volume sources so a misconfigured back_to / as_of window cannot pin query workers. If a query exceeds the cap, the caller gets the same class of descriptive error. See Fine-grained permissions.

​
Dataset health cards and symbol discovery

The dataset catalog now surfaces Concordance-backed health on each dataset:
  • Date range covered
  • Symbol count
  • Status (for example, Ready)
  • Last updated
The same fields are included when you call liberator.datasets(schema=True, details=True). Administrators can configure how often symbol discovery refreshes, so newly added keys appear on cards without waiting for a fixed backend job. See Listing datasets and Checking dataset access.

​
Excel plugin in the Liberator UI

The CloudQuant Excel plugin is available from the Liberator client experience. Users can find and download it in the UI instead of waiting on a separately shared installer. Account setup (User ID, token, Base URL) is unchanged. See Configuring your account.

​
User onboarding

  • Combined create-and-enroll — Super Admins create a user and add them to Liberator permissions in one step, instead of Create User followed by Add Users.
  • Username in invite and verification email — new-account mail includes the username the recipient should sign in with.
  • Search by company — the Users list finds people by their company field.
See Managing users and entitlements.

​
Per-query usage log

Super Admins can query usage_log, a private Liberator dataset of one row per query per dataset. Each row records the user, the dataset they asked for, the as_of / back_to they passed, rows returned, and an outcome of success, limit_exceeded, not_entitled, or error. Query it with the standard client (liberator.query(name='usage_log', ...)). The log’s own back_to / as_of select when the usage happened; the as_of / back_to columns are the window the original caller requested. The symbol / key is the dataset that was queried. Records are retained for at least 30 days. See Query usage log.

​
Connection and query reliability

  • S3 request style — during Test Connection, Liberator detects whether the endpoint needs path-style or virtual-hosted addressing (including MinIO, Ceph, and other S3-compatible stores), so you no longer have to set request_style by provider. See Amazon S3.
  • Database port is required on Test Connection for PostgreSQL, MySQL, SQL Server, and Oracle.
  • Passwords with special characters (%, ?, #, @, /, :, &) connect successfully on database sources.
  • Fixed / manual timestamps — datasets onboarded with a fixed date-time (no date column) query the correct partition after discovery.
  • Time-range queries resolve partitions more reliably across the requested window.
  • Future as_of on live-append datasets no longer holds query workers in an empty-fetch loop. Prefer as_of at or before “now” unless you intentionally query forward-dated records. See Usage of the as_of parameter.
  • HubSpot rate limits — HubSpot-backed queries wait and retry when the CRM API returns a rate-limit error. See HubSpot.
  • Date-format ranking — onboarding suggestions prefer ordinary calendar formats over day-of-year when the column contains month/day/year values.

​
Compatibility

There are no breaking changes to the client libraries (Python, REST, JavaScript, Excel, etc.) or to the dataset query surface. Existing exact-symbol queries continue to work. Pattern filters, day-range caps, catalog-card fields, and usage_log are additive.