Skip to content

feat(connectors): add the AutonLab TimeSeriesExam connector - #799

Merged
masquare merged 8 commits into
feature/label-set-derived-namefrom
feature/tnet-175-add-the-autonlabtimeseriesexam1-connector
Nov 20, 2026
Merged

masquare merged 8 commits into
feature/label-set-derived-namefrom
feature/tnet-175-add-the-autonlabtimeseriesexam1-connector

Conversation

@masquare

@masquare masquare commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Add autonlab/timeseriesexam1: the multiple-choice questions of AutonLab/TimeSeriesExam1 (MIT, arXiv:2410.14752). The connector pins the Hub commit 628f686757bb484d9b608b6855fb41f6a302ecda.

What it builds

  • One Record per series. A question has one series (ts) or two (ts1 and ts2). Each is one Signal on an ordinal axis, float64 and dimensionless. The values are not changed, and series keep their published lengths.
  • One ClassificationTask per question, in the published test split. The prompt is the question. The labels are the option texts, and the target is the published answer.
  • The other published columns stay in the task metadata, as published.
  • The dataset card has the licence, the citation and the source link.

Notes

  • The connector reads the rows through the central ParquetDataReader. It checks only that a row has ts, or ts1 and ts2. TimeNet checks the options, the answer and the series values.
  • The reader turns the series into Python lists while it converts. That costs about 0.15 s for this file.
  • The prompt does not list the options. The label sets are not ordered: the same options are in a different order in different rows, so the order has no meaning.

Verified against the real release

A build at the pinned revision gives 996 records and 746 tasks. The writer stores 148 label sets, and verify() passes.

Linear: https://linear.app/aionic/issue/TNET-175/add-the-autonlabtimeseriesexam1-connector

@masquare
masquare added this pull request to stack #800 November 18, 2026 01:36
@masquare masquare added the area:connectors Changes packages/timenet-connectors label Nov 18, 2026
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch 2 times, most recently from 7264ffb to 9c2087a Compare November 18, 2026 02:01
@masquare
masquare removed this pull request from stack #800 November 18, 2026 02:18
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from baf0779 to f039e63 Compare November 18, 2026 02:18
@masquare
masquare changed the base branch from chore/arfbench-label-set-names to feature/parquet-arrow-columns November 18, 2026 02:18
@masquare
masquare added this pull request to stack #803 November 18, 2026 02:18
@masquare
masquare removed this pull request from stack #803 November 18, 2026 02:20
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from f039e63 to fb85b85 Compare November 18, 2026 02:20
@masquare
masquare changed the base branch from feature/parquet-arrow-columns to chore/arfbench-label-set-names November 18, 2026 02:20
@masquare
masquare added this pull request to stack #804 November 18, 2026 02:20
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch 8 times, most recently from 7a00ca2 to 501a64a Compare November 18, 2026 07:06
@masquare
masquare requested a review from karhu November 18, 2026 07:08
@masquare
masquare marked this pull request as ready for review November 18, 2026 07:08
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch 2 times, most recently from d80f51a to 6bd4db7 Compare November 18, 2026 22:03
@masquare
masquare removed this pull request from stack #804 November 18, 2026 23:46
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from 6bd4db7 to ac5a07c Compare November 18, 2026 23:46
@masquare
masquare added this pull request to stack #813 November 18, 2026 23:46
@github-actions

github-actions Bot commented Nov 18, 2026 •

Copy link
Copy Markdown

✅ CI checks passed

Check Status
Quick checks ✅ Passed
Tests (Python 3.13) ✅ Passed
Tests (Python 3.11) ✅ Passed

Updated for 747e55d · run #8647

@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from ac5a07c to efa6cca Compare November 19, 2026 01:42
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch 3 times, most recently from 1f13dfe to bf9de6f Compare November 19, 2026 06:21
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from bf9de6f to 3a45a38 Compare November 19, 2026 16:17
@masquare
masquare removed this pull request from stack #816 November 19, 2026 16:18
@masquare
masquare added this pull request to stack #824 November 19, 2026 16:18
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from 3a45a38 to 75a252f Compare November 19, 2026 17:34
@masquare
masquare removed this pull request from stack #824 November 19, 2026 17:34
@masquare
masquare added this pull request to stack #826 November 19, 2026 17:34
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from 75a252f to 7dd632e Compare November 19, 2026 17:42
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from 7dd632e to 43726dd Compare November 19, 2026 17:48
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from 43726dd to 3555286 Compare November 20, 2026 07:25
@masquare
masquare removed this pull request from stack #826 November 20, 2026 07:25
@masquare
masquare added this pull request to stack #829 November 20, 2026 07:25
Add autonlab/timeseriesexam1, the multiple-choice questions of
AutonLab/TimeSeriesExam1 (MIT, arXiv:2410.14752), pinned to one Hub commit.

- One Record per series (ts, or ts1 and ts2), each with a lazy Signal on an
  ordinal axis. The values stay in the parquet file until the writer reads
  them.
- One ClassificationTask per question, in the published test split. The
  labels are the option texts, TimeNet names the label set from them, and
  the target is the published answer. The prompt is the question alone.
- The other published columns stay in the task metadata, as published.
- ParquetDataReader reads the rows, and the connector adds each row to the
  dataset as it reads it. The connector checks only that a row has ts, or
  ts1 and ts2. TimeNet checks the options, the answer and the series values.

The real release builds 996 records, 746 tasks and 148 label sets, and
verify() passes.
The options of a question are an ordered label set, so each question
keeps its published option order and the release stores one set for
each distinct order: 313 sets for the 746 questions.
Drop the tests that other tests in the file already cover, and the ones
that test LabelSet and writer behaviour of the core package. The three
reject tests are one parametrized test.
The source row is plain data again. convert builds one Record per
series and the task of each row. The option label sets are unordered:
the same options are in a different order in different rows, so the
order has no meaning.
TimeFDataset.write returns the version directory, so the test does not
repeat the dataset id and version. The pin test repeated the class
constants and is removed.
@masquare
masquare force-pushed the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch from 3555286 to 747e55d Compare November 20, 2026 15:31
@masquare
masquare merged commit 1bdc887 into main Nov 20, 2026
4 of 8 checks passed
@masquare
masquare deleted the feature/tnet-175-add-the-autonlabtimeseriesexam1-connector branch November 20, 2026 18:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

area:connectors Changes packages/timenet-connectors

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants