WSL gui_run.py not able to create needed files #17
-
|
Hey, I'm completely new to python or any command line stuff, but I can sorta follow directions. I have fought with the installation process and managed to get gui_run.py to eventually convert an EPUB to an audiobook, but when I tried to do another book it couldn't clear the previous book out of its system. I uninstalled and reinstalled everything, but now I am back to getting errors that the program can't find files that it hasn't created yet. What am I doing wrong here? |
Beta Was this translation helpful? Give feedback.
Answered by
DrewThomasson
May 9, 2024
Replies: 3 comments 25 replies
-
|
What OS are you running this on? And did you install calibre as in the guide |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
Windows 11 v 22H2. And yes, calibre ( 5.37) is installed. Here's all the programs in the environment: |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
Hm... how did you install it? Did you use the auto Ubuntu install command within the WSL Ubuntu machine? Or the manual process? |
Beta Was this translation helpful? Give feedback.
25 replies
-
|
Yup works on my end with colab I'll see if I can also get a google colab headless python script that uses the google colab easy "gui" type things to make it easier But let me know if it still doesn't work on your end for colab :/ |
Beta Was this translation helpful? Give feedback.
-
|
Whomp whomp. Still no dice. Exact same issue. Not sure what I can augment on my end so that we get similar results? It accepts the "book" and starts parsing it but falls apart after a minute or so. |
Beta Was this translation helpful? Give feedback.
-
|
You completly reset the runtime and such in the colab before running again right? Hm |
Beta Was this translation helpful? Give feedback.
-
|
Good news, tested with the file you were using.. There's something weird about the test pdf you're giving it.... Idk calibre won't convert it it spits out a blank file potential fixTry giving the program a txt file as a test instead EPUBs seem to also work Yeah pdfs are just weird, |
Beta Was this translation helpful? Give feedback.
Answer selected by
DrewThomasson
-
|
Whaaaat. You're not wrong. Damnit. I've tested it a few different ways and been able to get it to work. Thank you so much for all your help with this! I can't get EPUBs to work consistently. But txt has no problem. PDFs I'm not sure I've gotten to successfully work at all yet. Do you know if any of the programs being used to interrogate the EPUB ignores things like footnotes and picture text (e.g., Figure 1.1 "This is a picture of an extremely ancient opossum")? I'm trying to convert academic books to audio, but they're just lousy with footnotes and figures. I know I can remove them myself, but I'm not sure if that's contributing to the EPUB inconsistency. |
Beta Was this translation helpful? Give feedback.
-
|
lol your good, at least we got something working! About the images things? Honestly, I'm not sure, The program uses Calibre to convert everything to epub and then from epub to txt, so... yeah idk you'll have to test with Calibre on your end for that.. I think they have a windows app for calibre.... but other than the side note of epub2txt which isn't really used anyway, I think I used a custom beautiful script to detect chapters tho but yeah. idk |
Beta Was this translation helpful? Give feedback.
-
|
Hm. What is the benefit of going from epub to txt? Does it provide some kind of structure for the program? I was wondering about which epubs it's happy with since I can do a validation check with Calibre and it'll be happy, but your program won't like the cut of its jib for some reason. I'm wondering if I should cut out the middle man (headache) and just convert from EPUB to txt before running it? Do you have any information here about the chapter detection? It's been hit and miss. In the gui it has that "Silence duration" box, but I'm not sure if it's generating silence for chapters or looking for that duration for chapter breaks. If the latter, presumably that is for if there is no TOC? |
Beta Was this translation helpful? Give feedback.
-
|
Oh yeah I need to take out silence duration thing and chapter eliminator fields they arn't really doing anything right now The silence duration was suppose to change how long of a pause there is between each chunk of audio combined. So like if I didn't do that then there would be no pause between people talking and it would sound weird For chapter detection I had to make a custom script using beutiful soup, idk I'll have to look at it again but it works best for EPUBs. I tried with Calibre but it messed up more than my custom script did Also the reason it converts everything to epub and then to txt is cause it provides structure as you guessed, cause the chapter detection only works for EPUBs I just convert everything to EPUBs hope for the best, then convert rhat to txt to throw into BOOKNLP for quotation attribution and identification. |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Good news, tested with the file you were using..
There's something weird about the test pdf you're giving it....
Idk calibre won't convert it it spits out a blank file
potential fix
Try giving the program a txt file as a test instead
EPUBs seem to also work
Yeah pdfs are just weird,
I'll put a warning in the readme should of expected this lol