mmorrisj/Tamasheq-English-French
Domain:
natural language processingRecord type:
dataset
Creator:
mmo
Host:
Small corpus containing a dataset between Tamasheq (TAQ), English (ENG), and French (FR)
# Tamasheq-English-French
***
## Contributors
Cécile Macaire, Tomáš Pavliček, Alejandro Ciuba, Harshita Diddee, Marek Sarvaš, Santosh Kesiraju
Made available by Alejandro Ciuba, alejandrociuba@gmail.com
***
## NOTICE
**THIS DATA HAS BEEN COLLECTED FROM VARIOUS RESOURCES. THE USE OF THIS DATA IS ONLY FOR ACADEMIC NON-COMMERCIAL PURPOSES.**
***
## Summary
Small corpus containing a dataset between Tamasheq (TAQ), English (ENG), and French (FR). The data is organized amongst the following directories:
- `bible-data/`: **WARNING: DATA CONTAINS HEAVY BIAS!** Contains `bible-tamasheq.zip` and `bible-tamahaq.zip` (language similar to TAQ).
- `bilingual-dictionary/`: Contains files pertaining to our bilingual dictionary extracted from source text\*, as well as `bilingual-dictionary.zip`.
- `parallel-sentences/`: Contains files pertaining to our parallel sentences extracted from source text\*, as well as `parallel-sentences.zip`.
- `taq-fr/`: Contains files pertaining to our bilingual TAQ-FR dictionary, as well as `taq-fr.zip`.
***
## Directory Layouts
- `bilingual-dictionary/`:
`taq-biling.txt`:
`tamasheq_sing., tamasheq_plural (opt.)`
`eng-biling.txt`:
`eng_sing., eng_plural (opt.) / secondary_translation (opt.)`
`full-biling.txt`:
| English Word(s) | Tamasheq Sing. | Tamasheq Plural (opt.) |
| :--- | :---: | ---: |
| parent / ancestor | ămaraw | imarawăn |
- `parallel-sentences/`:
`parallel-sents-dialects.txt`:
| Dialect (Ša and/or Za) | Tamasheq Sentence | English Translation (word\_1/word\_2) |
| :--- | :---: | ---: |
| Ša | ma dăr tǝnsed? | How did you sleep? |
| Ša | ma šănšed? | What are you buying/selling? |
| Ša Za | ma tǝkned? | What are you making/repairing? |
**NOTE:** Another copy `parallel-sents.txt` exists with the first column removed (no dialectal information).
- `taq-fr/`: Both dictionaries follow the same format, with the TAQ and FR columns switched to reflect filename-order.
`taq-fr.txt` and `fr-taq.txt`:
`word translated_word`
**NOTE:** …