WebPlotDigitizer is giving mirrored results

[BerendAtEekels]

If I run an automatic extraction it gives "mirrored"results:

[nbehrnd]

What do you mean by "mirrored results"? Mirrored in respect to the abscissa, the ordinate? How do you proceed exactly? Can you replicate e.g. the extraction of the green trace, fourth while counting from right to left, which (replot with gnuplot) then appears to be quite similar?

test.zip

[ankitrohatgi]

Out of curiosity, did you rotate the image while making the mask by any chance? I am aware of an open bug that could have possibly caused this: #303

[nbehrnd]

@ankitrohatgi I did not apply a data transformation. What is a bit annoying, however is an other thing: while it is possible to set e.g. a fixed precision and a column separator for the export of single traces digitized, e.g. fixed precision to 5 decimals in total and an explicit white space to separate the columns

is not reflected with the export via click "all datasets" -> export all data. For one, the columns are again separated by commas (one run with sed or awk and there are out of the picture), but the precision defaults back to many more decimals, than wanted:

Can future versions of the digitizer please present the data with the number of decimals set in the other (single trace) interface?

The observations refer to the application run remotely from the web address in Firefox (115.5.0esr (64-bit)), the test digitization of all 10 traces attached in the .zip archive below.

2023-12-08_multi_trace_problem.zip

[BerendAtEekels]

Dear Sir/Madam Below is what I do, but now after pressing Run, nothing happens at all. I am working on Chrome. Yesterday I only once succeeded to get a decent set of numbers from the purple line most left. Other lines only gave me strange sets of “datapoints”. I did not rotate or perform other operations on the image. I also send you the file with the “source image”. It is on the first page. KR Berend Evenblij ***@***.*** Van: nbehrnd ***@***.***> Verzonden: vrijdag 8 december 2023 18:38 Aan: ankitrohatgi/WebPlotDigitizer ***@***.***> CC: Evenblij, Berend ***@***.***>; Author ***@***.***> Onderwerp: Re: [ankitrohatgi/WebPlotDigitizer] WebPlotDigitizer is giving mirrored results (Issue #310) U ontvangt niet vaak e-mail van ***@***.*** Meer informatie over waarom dit belangrijk is<https://aka.ms/LearnAboutSenderIdentification> What do you mean by "mirrored results"? How do you proceed exactly? Can you replicate e.g. the extraction of the green trace, fourth while counting from right to left, which (replot with gnuplot) then appears to be quite similar? test.png (view on web)<https://github.com/ankitrohatgi/WebPlotDigitizer/assets/19709084/b19ade21-6f8d-41c9-8b75-97a04623bd72> test.zip<https://github.com/ankitrohatgi/WebPlotDigitizer/files/13619258/test.zip> — Reply to this email directly, view it on GitHub<#310 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BERVF7O4DTHAIHITEWZGWZTYINF6JAVCNFSM6AAAAABAM5IGLKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNBXGU3TIMRWG4>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[BerendAtEekels]

Dear Sir/Madam, Here is another attempt to replicate the left-most purple line. The results are in the attachment. Now I run the program locally on my laptop. ***@***.*** When I plot the data I get this: ***@***.*** KR Berend Evenblij Van: nbehrnd ***@***.***> Verzonden: vrijdag 8 december 2023 18:38 Aan: ankitrohatgi/WebPlotDigitizer ***@***.***> CC: Evenblij, Berend ***@***.***>; Author ***@***.***> Onderwerp: Re: [ankitrohatgi/WebPlotDigitizer] WebPlotDigitizer is giving mirrored results (Issue #310) U ontvangt niet vaak e-mail van ***@***.*** Meer informatie over waarom dit belangrijk is<https://aka.ms/LearnAboutSenderIdentification> What do you mean by "mirrored results"? How do you proceed exactly? Can you replicate e.g. the extraction of the green trace, fourth while counting from right to left, which (replot with gnuplot) then appears to be quite similar? test.png (view on web)<https://github.com/ankitrohatgi/WebPlotDigitizer/assets/19709084/b19ade21-6f8d-41c9-8b75-97a04623bd72> test.zip<https://github.com/ankitrohatgi/WebPlotDigitizer/files/13619258/test.zip> — Reply to this email directly, view it on GitHub<#310 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BERVF7O4DTHAIHITEWZGWZTYINF6JAVCNFSM6AAAAABAM5IGLKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNBXGU3TIMRWG4>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[nbehrnd]

@BerendAtEekels A few notes:

• The archive test.zip as filed December 8th contains a brief silent test.mp4 video how I typically digitize the data using Firefox (at present, version 115.5.0esr (64-bit)) with JavaScript permitted. So far, I did not try the digitizer with the Chrome browser.
• If you want to reply with an email to an issue report here, it is not necessary to include the whole unchanged block of the message you reply to.
• If you reply to a conversation/issue report here on GitHub, email is valid only if your reply is text only. Any form of attachment will be removed. On the other hand, to substantiate your argument for instance with a .png, a .zip archive ; this only is possible as an attachment in an open session in GitHub itself.
• Archive 2023-12-08_multi_trace_problem.zip includes file wpd_datasets.csv with all ten traces, labelled 01, 02,...09, 10 beginning with the first trace to the left (light purple, 50 A) digitized with DeltaX = DeltaY = 10 px step width.