separating content vs function words

These are the function words that made it to the list:

a, an, the, this, that, these, those, my, your, their, our, some, many, few, all, and, but, or, so, because, although, in, of, on, with, by, at, over, under, he, she, it, they, we, you, me, him, her, is, am, are, was, were, has, have, had, can, could, may, might, shall, should, will, would, must, who, what, when, where, why, how

The processing conditions are:

• Two or more words in the title (this prevents cases where the movie is just a function word in the list, such as "her").
• At least one word must be present after filtering.

(I'm home & coding! )

Okay! Regarding the steps on Reply #5:

This is being tackled in two steps/releases:

- First, the MovieList for direct movie comparison.

- Second, Fuzzy-matching for adding more/partial results.

Direct movie comparison is good enough, so, now we proceed to implement fuzzy matching on the pre-processed titles, as derived from this step

The library of choice is "FuzzySharp", which is based on the previously-mentioned FuzzyWuzzy, hence it is functionally equivalent for our purposes.

• https://github.com/JakeBayer/FuzzySharp

fuzzy match works!

document the default ratio / partial ratio / token set etc from fuzzysharp into help menu?

alpha-4 has a weird bug where the matches result lists switch. its not important bug. i just report it.

Anthropophagous
First list:
Anthropophagous.1980.REMASTERED.1080p.BluRay.x264.DTS-FGT.mkv.torrent
Anthropophagus (1980) DTS (Custom 5.1 Mix).torrent
Anthropophagus 1980 1080p GBR Blu-ray AVC LPCM 2.0-CultFilmsT.torrent
Anthropophagus.1980.1080p.BluRay.x264-FAPCAVE.torrent
Anthropophagus.1980.BD.Baggerinc.torrent
ANTROPOPHAGUS.Mg.torrent
Antropophagus.torrent
Second list:
Anthropophagous [1980]

fuzzy matching is slower. i know there is nothing that can be done to speed this up, no worries. a progress/working indicator or progress bar would be nice. maybe just a little animated gif ?

program crashes on more characters:

Dark Lady of Kung Fu µûŸoÓ° (1981) º£°¶°æ.avi.torrent

Dark Lady of Kung Fu µûŸoÓ° (1981) º£°¶°æ.avi.torrent

You're correct: pre-processing throws an error when processing such odd titles.

The easy way out would be to strip non-alphanumeric characters in the traditional file-friendly way ("[^A-Za-z0-9_. ]+") but it can lose some title-delimiter information in the process.

I'm fixing it by keeping the processing "as is" for the correct file names while stripping bit by bit when reaching the ones that throw the error.

Wouldn't that all be fixed by handling strings as Unicode?

Hello & good day! You're right with your spot-on Unicode observation... yet, in this case, this particular app needs to have the lesser character pool to compare words in the movie title; we are currently folding characters to ASCII in order to help keep the program's complexity reasonable -- according to our standards, that is

Certainly, other implementations of fuzzy movie file name comparison can do it differently; it's just that we are really streamlining things here for our purposes... The lesser, the better in this case

another string that causes crash.

"FÃ¥rö Document    [Documentary]  [1970].mkv"

can we get an update program that doesn't crash on these?

It's solved! In the end the pre-processing is made with the regex "[^\u0000-\u007F]+" and then stripping all invalid path characters.

This approach solved everything regarding crashing due to weird symbols, in exchange of folding the foreign characters on the display title --of course, the actual passed file name remains intact.

I'm bumping the version to 0.2.0 to call main functionality complete

(Going to push to GitHub as time permits later today/tomorrow)

thanks vic

My pleasure 🤗

Please feel free to post a sorted list with the next functionality you need, in the order you want it developed

e.g. progress bar, auto saving, drag and drop, command line interface... plus any other.

(simply place those you consider the most important ones to implement at the top of the list)



Meanwhile, I continue "tweaking" this version's code to release

Cheers!

Hello & good day! New v0.2.0 is released:

• Download: MovieList-compn_v0-2-0.zip
• GitHub tag: https://github.com/paradisusis/movielist-compn/releases/tag/v0.2.0

1. command line interface ...

...On the way!

Cheers!

no more crashing!
also it doesnt use a lot of ram,  even on large lists. i am impressed!

Glad you're liking it!

In order to have separation of concerns, the new CLI version has been separated to its own thread and code:

https://www.donationcoder.com/forum/index.php?topic=54226.0

The current Windows/GUI version continues on v0.3.0 with drag&drop on the text boxes for the lists + other features, under its own code repository, as seen above

still run into annoying stuff like this with fuzzy match though

Bruiser
First list:
Bruiser
Gunbuster (1988)
buster.1988.internal.bdrip.x264-ghouls.mkv
Second list:
Buster (1988).mkv.torrent

Hello & good day! You may want to try with "Weighted" as algorithm and using different "cutoff" values.

Cutoff defaults to 75. In your case, you may want to try with 80, then 85, then 90 up to 95.

(Until you find the configuration that behaves closest to what you need/expect)

Cheers!