blob: 9b8da9c48b718ddbb27018ed5996636467aa96e8 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
|
# Evaluating cross-lingual textual similarity on dictionary alignment
This repository contains the scripts to prepare the resources for the study as well as open source implementations of the methods.
## Requirements
- Python 3
- nltk
```python
import nltk
nltk.download('wordnet')
```
- [lapjv](https://pypi.org/project/lapjv/)
- [POT](https://pypi.org/project/POT/)
- [mosestokenizer](https://pypi.org/project/mosestokenizer/)
- (Optional) If using VecMap
* NumPy
* SciPy
<details><summary>We recommend using a virtual environment</summary>
<p>
In order to create a [virtual environment](https://docs.python.org/3/library/venv.html#venv-def) that resides in a directory `.env` under home;
```bash
cd ~
mkdir -p .env && cd .env
python -m venv evaluating
source ~/.env/evaluating/bin/activate
```
Inside the virtual environment, the python interpreter and the installed packages are isolated.
In order to install all dependencies automatically;
```bash
pip install -r requirements.txt
```
After done with the environment run;
```bash
deactivate
```
</p>
</details>
## Acquiring The Data
```bash
git clone https://github.com/yigitsever/Evaluating-Dictionary-Alignment.git && cd Evaluating-Dictionary-Alignment
./get_data.sh
```
This will create two directories; `dictionaries` and `wordnets`.
Linewise aligned definition files are in `wordnets/ready`.
## Acquiring The Embeddings
We use [VecMap](https://github.com/artetxem/vecmap) on [fastText](https://fasttext.cc/) embeddings.
You can skip this step if you are providing your own polylingual embeddings.
Otherwise;
* initialize and update the VecMap submodule;
```bash
git submodule init && git submodule update
```
* make sure `./get_data` is already run and `dictionaries` directory is present.
* run;
```bash
./get_embeddings.sh
```
Bear in mind that this will require around 50 GB free space.
|
