sweagent-traj

name	sweagent_traj
description	Parse classic SWE-agent .traj trajectories: one step per entry of the trajectory array, in order, matching the CodeTraceBench annotation convention. No upstream CodeTracer commit ships a SWE-agent parser.
fingerprints	["*.traj"]
priority	60
metadata	{"version":"1.0","source":"agent-eval codetracebench-oht2-20260801"}

SWE-agent .traj Parser

Directory Layout

trial_dir/
  <trial>.traj             # JSON: environment, trajectory[], history[], info, context
  <trial>.context.json     # edited/viewed file context (not steps)
  <trial>.patch            # final submission patch
  report.json              # evaluation outcome

The run directory must hold exactly one *.traj file; zero or several refuse detection instead of guessing.

Step Convention

• One step per trajectory[] entry, in array order. Every entry counts, including entries whose action is blank.
• The action is the entry's raw action; actions that are blank after trimming render as their JSON literal so the step survives non-empty action gates losslessly.
• The entry's observation passes through unchanged (empty strings stay empty strings); a non-string, non-null observation is a parse error.
• The entry's thought becomes thinking when it is a non-blank string.
• action_ref points at <trial>.traj#trajectory[i] and embeds the raw entry JSON (action, observation, response, state, thought), so the full source record stays recoverable per step; observation_ref stays null because it would duplicate the same entry.
• The task description is the first user history[] message that is not the interface demonstration (marker: --- DEMONSTRATION ---). Both published prompt templates (<pr_description> upload and ISSUE: text) sit in that message on all 108 verified rows.

This enumeration reproduces the CodeTraceBench annotation step_count on 106 of 108 verified SWE-agent rows at dataset revision aa213b84ffb6690fc37ca15766d6ca174ec36d4d. The other 2 rows (keras-team__keras-19775, mui__material-ui-11451) publish a function-call-style trajectory (45/48 and 92/94 entries with empty response) whose annotated view condensed the run (26 vs 48, 54 vs 94 entries); no published view reproduces those counts, so they fail the step-count gate loudly instead of importing misaligned labels.