perda.timeline.client#

class perda.timeline.client.TimelineClient(password=None)[source]#

Bases: object

Parameters:

password (str | None)

find(var_key, above=None, below=None, month=None, min_samples=1)[source]#

Find sessions where a variable crossed a threshold. Answered entirely from the summary tier – no raw samples are read.

Parameters:

  • var_key (str) – Exact C++ identifier for the variable.

  • above (float | None) – Keep sessions whose maximum exceeded this.

  • below (float | None) – Keep sessions whose minimum fell under this.

  • month (str | None) – Restrict to a month, formatted "YYYY-MM".

  • min_samples (int) – Ignore sessions with fewer samples than this.

Returns:

Matching sessions, worst-first.

Return type:

pl.DataFrame

load(var_key, session_id=None, test_day=None)[source]#

Load a variable as a PERDA DataInstance for plotting and maths.

Bridges the timeline back into the existing analysis stack, so a signal pulled from months of history behaves exactly like one parsed from a single log.

Parameters:

  • var_key (str) – Exact dotted C++ path.

  • session_id (int | None) – Restrict to a single session.

  • test_day (date | None) – Restrict to one test day.

Returns:

Timestamps in microseconds relative to the session start.

Return type:

DataInstance

overview()[source]#
Return type:

DataFrame

samples(var_key, session_id=None, test_day=None)[source]#

Fetch raw samples for one variable.

Parameters:

  • var_key (str) – Exact dotted C++ path.

  • session_id (int | None) – Restrict to a single session.

  • test_day (date | None) – Restrict to one test day.

Returns:

Columns session_id, ts_utc, t_rel_us and value.

Return type:

pl.DataFrame

sizes()[source]#
Return type:

DataFrame

sql(query, params=None, max_rows=1000000)[source]#
Return type:

DataFrame

Parameters:

  • query (str)

  • params (Sequence[object] | None)

  • max_rows (int)

stats(var_key, month=None, test_day=None)[source]#

Per-session summary rows for one variable.

Parameters:

  • var_key (str) – Exact C++ identifier for the variable.

  • month (str | None) – Restrict to a month, formatted "YYYY-MM".

  • test_day (date | None) – Restrict to a single test day.

Returns:

One row per session, ordered by start time.

Return type:

pl.DataFrame

var_id(var_key)[source]#

Resolve a variable key to its canonical id, with a local cache.

Parameters:

var_key (str) – Exact dotted C++ path.

Returns:

The canonical var_id, or None if the key is not catalogued.

Return type:

int | None