Mark : Thanks for the reply. I just needed to use a bit more common sense on this one. I wanted to change the curve name and I actually looked at the Add/Remove Curves menu and thought it wasn't what I what I wanted because it didn't have an Edit button. Which is my mistake. All I have to do is double click on the name and I can change it. Thank you.
Hi, terribly sorry to ask such a niave question, but how does one install engauge-digitizer from either a .zip or .tar.gz file? Please don't point me to the INSTALL .md file because telling me to download binaries does me no good since I don't know what those are and I haven't found instructions for newbies. I'm using Windows 10. Thank you!
Good question. The zip and gz files are 'compressed' files containing multiple files stored inside. To uncompress them you typically right click on them and select an option like Extract or maybe 7z which is a particular tool for uncompressing them. The result of uncompressing is usually a new directory in the same directory as the compressed file (unless you specified somewhere else during uncompression).
In Windows 10 you right click on the zip file. Click on Extract All. In the Extract Compressed(Zipped) Folders dialog enter the directory where you want the new files. Click on Extract. Good luck!
Oh, and double click on engauge.exe in the bin directory to run it
Thanks for your help, @markummitchell! I should have been clearer, though: I do know how to unzip files, but the only executable is not in the bin folder and it's not called engauge.exe. There's an executable called "nasm-2.14rc0-installer-x64.exe" in dev/windows/poppler-tarballs, but it makes me nervous to install an executable that's not where I was expecting nor called what I was expecting. Is that what I should be using?
I can look into this in about 3 hours
newly->newby
Took a little longer to respond than expected. Computer died so I had to replace it. Regarding yur problem, it looks like Windows Defender now deletes some exe files. I verified that the exe files are in the v10.77 release. You can follow the instructions at https://answers.microsoft.com/en-us/windows/forum/windows_8-security/how-to-stop-windows-defender-from-deleting-a/f4e3641a-5598-495d-b2a5-bd6ead226071?auth=1., or use use an Engauge installer (*.msi). Some people on work computers are unable to use the installers since they do not have admin rights.
Hi Mark - thank you for maintaining such a great tool. We use it a lot in our day to day data capture work. Now I ran into a problem and was hoping that you could help. We want to use our digitized figures as a training set for an neural network application. We do have the exported data tables but did not always save the images we imported to Engauge. Do you see an easy was to export the images (e.g. in png) from an existing Engauge file?
Please ignore my previous post. I have added a new -extractimageonly command line flag that will work for any version of input file. After that argument goes an image file extension like 'png' to select the output image file format. You can use a Windows pre-release with this feature at https://ci.appveyor.com/project/markummitchell/engauge-digitizer/branch/master/job/ysmflr09mufpxf4m/artifacts
I love this tool!! However, I need to create a data file for an arbitrary waveform generator that requires 8192 horizontal data points. At the moment I am changing the "settings" to output one "y" value per pixel but that onlt gives me (say) 350 points. Is there a way to generate 8192 ???
You should be able to select Graph Units in the Export Format dialog, then just set the interval to a (xmax-xmin)/8192. Issue #308 may slightly complicate this because zero is used instead of xmin - you can either wait a few days or work around that yourself if that problem appears. Good luck
how to fix?
There is so much information to translate that I had to use machine translation for some languages. If you can help, that would be great. Even with machine translation, the copying and pasting can take two hours so you may want to translate just the text you find most important and/or needing the most help. The easiest approach seems to be with Transifex.com. Translators can get a free account and then request translation privileges for https://www.transifex.com/engauge-digitizer/engauge-digitizer/. There are 200 known downloads per day of this tool, with 11% from Germany (https://snapcraft.io/engauge-digitizer/metrics) so you would be helping a lot of people! I hope you are interested.
@kaufdev Sorry, but Engauge is not designed to work automatically, like 'engauge <image_file_in> <csv_file_out>'. The user interface right now has many, many features to handle the wide range of inputs that users work with, and the work to automate all of those features would take months at least. So just running a single command line and letting the software automatically digitize an image will not be possible. If you were just asking about using the current user interface in a text window, that would not be possible because the software is based on Digia's Qt which does not work in terminals. Sorry
@sharonrlin_gitlab This is in work as markummitchell/engauge-digitizer#315. You can try out a pre-release for 64 bit Windows at https://ci.appveyor.com/project/markummitchell/engauge-digitizer/branch/master/job/c9dlm24t20ulviur/artifacts
I'm having a problem with 11.1 and had to fall back to an older copy. What was happening was the image I was digitizing would move up/down slightly at the end of the curve. It was as if I clicked on the up/down arrows on the right side. I've zoomed in until my image is larger than the screen as there is white space and I wanted my image to digitize as large as possible. When this happened, the Y coordinates would be off by the amount the image shifted.
Would it be possible to have a magnifying glass around the region to make your selection more accurate when you are trying to follow a curve?
Okay, I'm back. What I've noticed and can't pin point how it happens is that if you have a JPG that you are scanning and the image file is larger than the display window, the image will shift up and down sometimes without hitting the arrow keys or rolling the mouse wheel. I can't figure out what triggers it but it causes the digitization to be incorrect.
fix
The problem to using Fit is that it changes the ratio of length to width and I haven't had time to test to see if that alters the digitized output. If it would fit to either hight or width and keep the same ratio that would be nice. I'll try to get more time to test it. Another thought might be to have the ability to lock the image from panning up or down along with the mouse wheel just in case.
Mark just in case you don't have anything else to do. Have you considered creating an option to store the X&Y values in the buffer so you can simply paste them directly into an Excel file rather than having to create a temporary Excel file, paste them, and them copy the columns to the destination file? I typically create one Excel file with all of the digitized data for one group of equipment rather than individual worksheets for each and to be able to just select a starting cell and have the program populate the cells would be neat.
@oopman2002 I searched the code and the only thing that should cause scrolling is when zooming in or out - in which case the window scrolls to try to keep the cursor position in the center of the window. Are you running any zoom utility? If not, I will probably need more information, like are you at the left or right edge when the image jumps up or down? Is a point selected?
@oopman2002 Regarding zooming, if you increase/decrease the zoom factors, using Control+MouseScroll, then you should get 1:1 aspect ratio
@markummitchell Mark, you helped me in the past when I needed a way to export the images from our 4000 Engauge files. We try to train an AI system with our digitized output and now have the images. We do have all the digitized points in the plot coordinates as well but would need the point position in "image-coordinates" - basically in pixel position from one of the image corners. Do you see a way to get that with not too much effort on your end? I assume that information is available somewhere before you do the coordinate system transformation with the identified axes points. A way to export those coordinates for each digitized point in our 4000+ files would be amazing. (We can script somthing in Python too, if you can maybe point us to where we should start looking).
@wilthan Issue #325 was created to track this. You can try out a new python script at https://github.com/markummitchell/engauge-digitizer/blob/master/contrib/python/DumpGraphAndScreenCoordinates.py. If you are comfortable with python, you can modify that script. Another option might be a script to read the DIG file and any CSV files created from it, and convert the points in the CSV files to create new CSV files. In any case, good luck with your AI system.