nomic-ai / gpt4all

GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.
https://nomic.ai/gpt4all
MIT License
68.98k stars 7.57k forks source link

[Feature] v3.0+ , Proper descriptions for the number of file(s) and word(s) #2586

Open SINAPSA-IC opened 2 months ago

SINAPSA-IC commented 2 months ago

Feature Request

On the LocalDocs page (and elsewhere where applicable), while Embedding, the number of files and words (like of any countable entities) should be properly described - for example, the number of files to be followed by the correct "files" when there are more than 1 file, and "file" when it's 1 file only.

Reason: When Indexing and Embedding the contents of files in collections of LocalDocs, the number of words (if not that of files) is larger than 1 - the image here shows such a case, when a number of +6 million words is followed by "word(s)", which is ridiculous however you spin it. The terms with "(s)", however familiar and truly nostalgic (introduced some 30 years ago in the first Windows programs - which means that this is a Professional approach, however anachronistic), in fact suggest to the user: 1) a lackadaisical approach to the simple display of a count, as if the programmer didn't care or know how many entities are there, even if they've just displayed/used that number, so they placed the "(s)" to cover all the bases 2) insecurity, as if that number may change unexpectedly from 1 to >1, which is not the case when the files and words due for Embedding were counted correctly before starting the procedure 3) lack of interest when displaying certain/known things within a program dealing with artificial intelligence - a far cry from nostalgia 4) this is not about aesthetics, but about certainty.

Suggestion: Since the number of files and words is known - as it was just being displayed ahead of the "file(s)/word(s)" - it is only logical that it should be followed by the associated term - "file" and "word" when there is only 1 file or word, and "files" and "words" when there are more than 1 file or word.

The image shows

SINAPSA-IC commented 2 months ago

img30_ui7

cebtenzzre commented 1 month ago

We are already using %n for this, see here. As soon as the English localization is working properly (I haven't checked recently), then this should show "file" or "files" automatically, without the need for a ternary.