Show sutra text #33
Show sutra text #33
Conversation
Generated with the following python script:
```
import requests, json
from indic_transliteration import sanscript
data = json.loads(requests.get('https://raw.githubusercontent.com/ashtadhyayi-com/data/master/sutraani/data.txt').text)['data']
print(len(data))
out = {}
for sutra in data:
name = f"{sutra['a']}.{sutra['p']}.{sutra['n']}"
text = sutra['s']
slp1 = sanscript.transliterate(text, sanscript.DEVANAGARI, sanscript.SLP1)
if slp1 == 'kftyErfRe':
back = sanscript.transliterate('kftyEr fRe', sanscript.SLP1, sanscript.DEVANAGARI).replace(' ', '')
elif slp1 == 'urft':
back = sanscript.transliterate('ur ft', sanscript.SLP1, sanscript.DEVANAGARI).replace(' ', '')
else:
back = sanscript.transliterate(slp1, sanscript.SLP1, sanscript.DEVANAGARI)
assert back == text, (text, slp1, back)
out[name] = slp1
with open('sutrapatha.json', 'w') as f:
json.dump(out, f, indent=2)
f.write('\n')
```
|
The async is kind of weird but it works, and loading Aside: note that transliterating to SLP1 makes it hard (AFAICT) to recover the original Devanagari (which is debatable in the first place) for:
— see indic-transliteration/indic_transliteration_py#75 (this is the kind of thing that I'm hoping a Rust transliteration library would fix by being "pedantic" and requiring specifying a strategy instead of making ad-hoc choices, but it's probably fine for now). |
|
Made a small change, updated screenshot (tested with manual removal of
|
Generated with:
```py
import requests, json, csv
from indic_transliteration import sanscript
data = json.loads(requests.get('https://raw.githubusercontent.com/ashtadhyayi-com/data/master/sutraani/data.txt').text)['data']
print(len(data))
out = []
for sutra in data:
name = f"{sutra['a']}.{sutra['p']}.{sutra['n']}"
text = sutra['s']
slp1 = sanscript.transliterate(text, sanscript.DEVANAGARI, sanscript.SLP1)
# if slp1 == 'kftyErfRe':
# back = sanscript.transliterate('kftyEr fRe', sanscript.SLP1, sanscript.DEVANAGARI).replace(' ', '')
# elif slp1 == 'urft':
# back = sanscript.transliterate('ur ft', sanscript.SLP1, sanscript.DEVANAGARI).replace(' ', '')
# else:
# back = sanscript.transliterate(slp1, sanscript.SLP1, sanscript.DEVANAGARI)
# assert back == text, (text, slp1, back)
out.append((name, slp1))
with open('sutrapatha.tsv', 'w', newline='') as f:
writer = csv.writer(f, dialect='excel-tab')
writer.writerows(out)
```
|
wonderful -- thank you!! |
I'd love to discuss this with you further now that we have a starter implementation (https://ambuda-org.github.io/vidyut-lipi/). |
Hacky code for #26 , but it actually seems to work:
Before merging I guess at minimum we should copy over https://raw.githubusercontent.com/ashtadhyayi-com/data/master/sutraani/data.txt so that it's hosted here — maybe convert it to a simple
data/sutrapatha.tsvas suggested at #26 (comment) that has entries like— but just pushing this commit for now as a savepoint for when I return to this later (not today). (Or if I don't return 😱 )
(Curious how much faster it will get when we change from array of 4000 items to JS object for faster lookup…)