Sign inSign up

mtaril/emtsv

By mtaril

•Updated about 1 year ago

e-magyar text processing system for hungarian language

Image
1

10K+

mtaril/emtsv repository overview

⁠README

Docker image for e-magyar⁠ Text Processing System (github repo⁠).

Available modules:

  1. tok: tokenizer (emToken⁠)
  2. morph: Morphological analyser and lemmatizer (emMorph⁠ and emLem⁠)
  3. pos: POS Tagger (emTag⁠)
  4. conv-morph: emMorph to Universal Dependency converter (depToolPy⁠)
  5. dep: Dependency parser (emDep⁠)
  6. cons: Constituent parser (emCons⁠)
  7. chunk: NP Chunker (emChunk⁠)
  8. ner: NER Tagger (emNer⁠)

Dependency of modules:

                                   +------------+    +-----+
                              +--->+ conv-morph +--->+ dep |
                              |    +------------+    +-----+
                              |
                              |    +------+
                              +--->+ cons |
+-----+    +-------+    +-----+    +------+
| tok +--->+ morph +--->+ pos |
+-----+    +-------+    +-----+    +-------+
                              +--->+ chunk |
                              |    +-------+
                              |
                              |    +-----+
                              +--->+ ner |
                                   +-----+

⁠Usage

⁠1 Server-Client
⁠1.1 Run docker container on server
docker pull mtaril/emtsv
docker run --rm -p <port-on-server>:5000 -d mtaril/emtsv
⁠1.2 Call emtsv from client

For choosing modules to run, enumerate module names separated with slash, e.g. call emtsv with curl⁠ for NER tagging:

curl -F "file=@<path-to-input-file>" http://your.server.com:<port-on-server>/tok/morph/pos/ner

Or with httpie⁠:

http -f POST http://your.server.com:<port-on-server>/tok/morph/pos/ner file@<path-to-input-file>

Or from python script:

import requests

with open('path/to/inputfile') as inp:
    response = requests.post('http://your.server.com:<port-on-server>/tok/morph/pos', files={'file':inp})
print(response.text)
⁠2 Locale
docker pull mtaril/emtsv
cat input.txt | docker run --rm -i mtaril/emtsv tok,morph,pos,ner

Tag summary

Content type

Image

Digest

sha256:bb2e56806…

Size

3.6 GB

Last updated

about 1 year ago

docker pull mtaril/emtsv