The evaluation report needs calibration error across fixed confidence bins. Implement the calculation so bin boundaries, empty bins and their weights follow the report's definition.
Build the implementation in Python under /app, with /app/run.sh as its entry point. Read one JSON request from standard input and write one JSON result to standard output. Input has positive integer bins and samples with probability in [0,1] and binary label. Bin index is min(bins-1,floor(probability*bins)); probability 1 belongs to the last bin. Return counts for every bin and ece, the sample-weighted absolute difference between average probability and average label per bin, rounded to six decimals. Empty samples yield zero. Reject invalid values.
Reject inputs that violate an explicit validation rule with exit status 2, no standard output, and {"error":"invalid_input"} on standard error. For valid input, exit with status 0. Keep the implementation reusable for new inputs; it must work offline and must not depend on the example data being the only data it receives.
Repair or replace the unfinished /app/run.sh and add source files under /app. Malformed JSON must use the same invalid-input response. Do not read /tests or /oracle from the application.