nmdanny / yap-wrapper

Python wrapper for ONLP YAP https://github.com/OnlpLab/yap

YAP-Wrapper

Python interface to Bar-Ilan University ONLP lab (https://nlp.biu.ac.il/~rtsarfaty/onlp) YAP (Yet another parser) https://github.com/OnlpLab/yap.

Yap is a free tool, If you make use of YAP for research, we would appreciate the following citation: https://aclweb.org/anthology/papers/Q/Q19/Q19-1003.bib

Alternatively you may use the SaaS server (free use), https://www.langndata.com/heb_parser/overview that case you need only one line of code: https://www.langndata.com/api/heb_parser?token=[YOUR TOKEN HERE]&data="גנן גידל דגן בגן"

Installing

1. Install YAP (instruction are on YAP page.... https://github.com/OnlpLab/yap)

2. After installing YAP, run it as HTTP server, by simply run "./yap api" from command line. Now Yap is running as HTTP serve on port 8000.

3. Install this project via a git dependency - if using pip(preferably within a virtualenv), do pip install git+https://github.com/amit-shkolnik/YAP-Wrapper.git

   Alternatively, clone this repository and then do pip install . within the repository's folder.

Running the examples

Clone this repository and make sure the package is installed within your Python environment. You should now be able to run python3 examples/yap_api.py or python3 examples/heb_tokenizer.py

If YAP is not installed locally, set the IP address on examples/yap_api.py main method.

The code return 6 elements:

• tokenized_text- string. The text tokenized.
• segmented_text - string. The text segmented.
• lemmas - string. The lemmas of the text.
• dep_tree - DataFrame. Dependency tree.
• md_lattice - DataFrame. Morphological analysis, as decided by YAP.
• ma_lattice - DataFrame. Morphological analysis inlcuding all possible lattices.

For example, if original text is: "בתוך עיניה הכחולות" Then output is:

• tokenized_text- "בתוך עיניה הכחולות"
• segmented_text - "בתוך עיניה ה כחולות"
• lemmas - "בתוך עין ה כחול"
• dep_tree -

• md_lattice -

• ma_lattice -

This code is fully free under Apaceh 2.0 (https://www.apache.org/licenses/LICENSE-2.0) (no guarantees!).