Added disambiguated text to texts folder along with source text

[pranavad]

Formatted the disambiguated text a bit.

[shardulc]

Hi @pranavad, git shows these files as binary files and not text files. (See the 'Files changed' tab for example.) How did you create the files? Is there an issue with encoding or something?

[pranavad]

I saved them as Unicode encoding in Notepad. I'll try and export it again in UTF-8 perhaps?

[shardulc]

Yup, do that.

[pranavad]

I think it's fine now. I re-encoded them in UTF-8, and now it isn't showing them as binary files

[pranavad]

Hey, for something like Bay of Bengal/Indian Ocean, is the disambiguation for "Indian"/"Bengal" an adjective, an hydronym or a toponym?

[pranavad]

Similarly, is Ocean/Bay here a common noun or a proper noun?

[pranavad]

For words such as "Soviet Union" or "World Cup" where each word can be interpreted as a common noun, how do I disambiguate it?

Languages and sports are "altres" proper nouns?

And lastly, things like Hindustani Music, would have "Hindustani" as a proper noun or adjective?

[pranavad]

@shardulc
Here are the words that the analyzer didn't recognize I couldn't manually disambiguate

प्रभावित

इत्यादि

हालाँकि

उभरी

सोवियत

पुरजोर

अलावा

वाली

हिन्दुस्तानी

[shardulc]

@pranavad Sorry for the late response.

For all the examples like "Bay of Bengal", "Soviet Union", etc. the short answer is that do it the way the English module analyzes them, since that module is mature now. The long answer, if you want to know how/why things are done, is that I'm not sure about the technicalities and you'll have to ask on the channel.

From your list, प्रभावित and हिन्दुस्तानी seem to be normal adjectives to me. The comment above applies to the rest.

[pranavad]

Hey, @shardulc, I completed the disambiguation for all the lemmatized words as well as fixed a few others.

[pranavad]

Is there anything else required for merging/task completion?

[shardulc]

Approved your task! I'll look into merging with perhaps some minor changes shortly.

[Kainatic]

@pranavad @shardulc
How to open morphological analyser? (steps please)
Checked the docs. Nothing related to my Q.
I'm doing the same task.

[pranavad]

Hey @Kainatic. I had trouble understanding this at first as well. I'll help you out on IRC, join the Apertium channel

[Kainatic]

On IRC
help

[pranavad]

@Kainatic Off memory, you have to install lttoolbox first of all and download apertium-hin.hin.dix to your home folder in the Virtual Box after installation [you'll find steps for this in the doc]. Then, run
$ lt-comp lr apertium-hin.hin.dix hin.analyser.bin

(The hin.analyser.bin is a name that you can change)

Then you can echo any text you want to analyse by
$ echo "text" | lt-proc hin.analyser.bin

Hope this helps.

[Kainatic]

Thanks

[Kainatic]

@pranavad How did you do it?
Like, did you separate and copy each sentence, or pass the text file, or something else?
Please tell me the next steps.
And how did you measure exactly 500 tokens?

[jonorthwash]

The best way to compile apertium-hin is to install apertium, clone apertium-hin, go into the cloned directory, run ./configure and when that's done make.

[Kainatic]

@jonorthwash Went into the cloned directory and ran './configure' but it said 'bash: ./configure: No such file or directory'
Also, read about make. There is already a Makefile.am present in the cloned directory.
Require help about running hin-tagger mode on Wiki.txt to disambiguate.

[pranavad]

Hey, @shardulc. Should I make another PR for the constraint grammar task, or is this fine for now?