indic-dict/stardict-sanskrit

Add correction links to cologne dicts

vvasuki opened this issue · 11 comments

Suppose I am reading kalpadruma dict entry for tulasii and notice some text with several typos: " पद्मपुराणमते जलन्धरपत्नीवृन्दालावण्यदर्श- नेन विष्णुर्मुग्धः । तन्मोहवारणार्थं देवा महा- देवशरणङ्गताः । महादेवस्तानुवाच मायामारा- धय । ते मायां तुष्टुयुः माया तानुवाच अहं रजः- " . If there were a link beneath the entry to take me directly to a github page, I would fork, do the correction and send a PR.

I did not understand the question properly.

When you say, "I am reading..", what exactly do you have in mind?

  1. Reading on some stardict viewer or
  2. Reading on Cologne website ?

In both the cases, the best way to help would be to make necessary corrections in the following repository in the following location.
https://github.com/sanskrit-lexicon/csl-orig/tree/master/v02 and create a PR.
We would be able to merge it very fast.

If you or anyone is interested in correction submission on mass scale, I would gladly help setting up things.

  1. Reading on some stardict viewer or

Given that I opened the issue in stardict-sanskrit - On a stardict viewer (mostly on my computer, where typing would not be a chore) .

Rather than at a "mass scale" by a single person, we should should shoot for "mass scale" by crowd-sourcing. Casual stardict users should be able to contribute. Currently stardict users basically fret and ignore the errors. https://raw.githubusercontent.com/sanskrit-lexicon/csl-orig/master/v02/skd/skd.txt is too huge - it should be split to one file per entry, and that file's github edit path should be linked from the corresponding stardict entry.

Casual stardict users should be able to contribute

Nice idea.

@vvasuki
Can you please test the following dictionary for the suggested features?
indic-dict/stardict-sanskrit/sa-head/en-entries/Meulenbeld-Sanskrit-Names-of-Plants

It seems to work fine on my android phone.

Looks beautiful -
image

Hope I'll be able to correct kalpadruma in devanAgarI though!

All regenerated.
Not sure about MW. All other dicts done. Kindly check.

Tried this with colordict on Android tablet (which I have not used for probably a couple of years).

First comment - was able to figure out how to update the dictionaries fairly easily, and all went smoothly.
Also getting information for an entry is straightforward.
The scanned image link worked perfectly.

Also tried Correction.

When pressed 'Correction submission: guru:7025' for Macdonell, I see a 'csl-lnum' Github page. But was shy about what to do next. If I really had a correction, what would I do to submit the correction? And what pipeline would the
submitted correction enter?

Incidentally, didn't see a correction submission link (or a scan page link) under MW dictionary for guru.

If I really had a correction, what would I do to submit the correction?

You would click on 'edit' button on csl-lnum page of guru of MD. Make necessary changes (if you have access rights to the repository).
If you do not have access rights, you would clone and create a PR.

MW dictionary for guru.

MW does not have this facility yet. I am yet to upload ig. Size of file exceeded 100 MB after addition of links. Will resolve and put to use soon.

And what pipeline would the submitted correction enter?

We can merge the PR, if correction is found OK.
I will write a script to carry changes from a given lnum of a given dictionary to csl-orig/v02/xxx/xxx.txt file. That will carry the change made at csl-lnum repo to csl-orig.
The idea of csl-lnum was to give a chewable bite size to csl-orig.

Alternative: As the corrections in csl-lnum by users are not going to be copious, we can use csl-lnum just as a reporting mechanism and carry changes to csl-orig on our computers and regenerate csl-lnum. That would serve both purposes of integrating user's correction and not accidentally damaging csl-orig while carrying some change back from csl-lnum to csl-orig.

The links have been added to csl-ldev repository.
This gives correction facility in Devanagari for Devanagari content.
The issue is closed now.