[ManuelOsburg]

Problem

The ConfigData docstring and the prompt for the image names recommend a format spec such as img{:04d}.jpg for the zero padded IDs most cameras write (DSC_0001.CR3). init_functions and DataExtractor cast the ID to int before formatting. ConfigData.get_start_time and ConfigData.in_time_diff_to_img_time pass the ID as read from config.ini, which is a string. With a format spec they fail:

ValueError: Unknown format code 'd' for object of type 'str'

get_start_time is called whenever start_time is None, i.e. it is taken from the first image, which is the default. So step 3 stops for every padded template unless start_time is set by hand. in_time_diff_to_img_time fails in the same way when the clock offset is computed from a time reference image given by time_ref_img_id.

Change

int(...) on the ID in both places, as in init_functions and DataExtractor.

For a plain {} template the name only changes if the ID in config.ini is written with leading zeros (0001). It is then img1.jpg instead of img0001.jpg. That is the name init_functions already uses for all experiment images, so this combination did not work before either, and no working setup changes.

Tests

New suite 09_test_image_names.robot. The test images are renamed to test_img_0001.jpg … and img_name_string = test_img_{:04d}.jpg is used. Two cases:

• the start time is taken from the first image;
• the clock offset comes from a time reference image given by its ID (time_ref_img_id = 1, time_ref_img_time = 12:00:00). The offset comes out as 0.0, computed from test_img_0001.jpg.

Both run S1–S3, check that the four padded images are analysed and that the linear extinction coefficients match the ground truth (rmse 0.014 / 0.016 / 0.022 < 0.05).

• Without the fix, on master, both cases stop with the ValueError above, each in its own function.
• With the fix, the whole suite passes 15/15, run locally on Windows without UTF-8 mode (cp1252) from a clean export of the commit. The encoding check of the CI is clean.
• Together with Feature/local background subtraction #131: the combined state of this PR and Feature/local background subtraction #131 also passes the whole suite. The two PRs do not conflict; the new keywords Rename Test Images To Padded IDs and Set Config Option sit apart from the lines Feature/local background subtraction #131 changes in LedsaATestLibrary.py. The suite is numbered 09, because Feature/local background subtraction #131 adds 08_test_background_subtraction.robot.

Observation, not changed here

The second test sets time_ref_img_id explicitly, because the config template writes the option as time_img_id, while the code reads time_ref_img_id. As a consequence:

• time_img_id set in config.ini, as the template suggests, is ignored. If exif_time_infront_real_time is not set, the user is asked for the ID instead.
• If the time reference image is given by time_ref_img_file and the config has no time_ref_img_id, in_time_diff_to_img_time raises KeyError: 'time_ref_img_id', because it indexes the option directly.

Both are the same on master; I reproduced them there. A fix would write time_ref_img_id into the template, read the option with a fallback, and still accept time_img_id from existing configs. I left it out to keep this PR small; happy to add it here or open a separate PR, whichever you prefer.