First load

.gitignore

@@ -15,6 +15,7 @@ dist/
 downloads/
 eggs/
 .eggs/
+.idea/
 lib/
 lib64/
 parts/

.travis.yml

@@ -0,0 +1,16 @@
+language: python
+python:
+  - "3.4.5"
+  - "3.5"
+  - "3.6"
+
+# command to install dependencies
+install:
+    - pip install -r requirements.txt
+    - pip install coveralls
+
+# command to run tests
+script:
+    - coverage run --source=pyperseus_treebank setup.py test
+after_success:
+    - coveralls

README.md

@@ -1 +1,4 @@
-# pyperseus-treebank
+# pyperseus-treebank
+
+[![Coverage Status](https://coveralls.io/repos/github/PonteIneptique/pyperseus-treebank/badge.svg?branch=master)](https://coveralls.io/github/PonteIneptique/pyperseus-treebank?branch=master)
+[![Build Status](https://travis-ci.org/PonteIneptique/pyperseus-treebank.svg?branch=master)](https://travis-ci.org/PonteIneptique/pyperseus-treebank)

pyperseus_treebank/base.py

@@ -0,0 +1,21 @@
+class Token:
+    def __init__(self, index, form, lemma, parent=0, features="", pos=None, rel="ROOT"):
+        self.index = index
+        self.form = form
+        self.lemma = lemma
+        self.parent = parent
+        self.pos = pos
+        self.rel = rel
+        self.features = self.parse_features(features)
+
+    def parse_features(self, features):
+        """ Parse features from the POSTAG of Perseus Latin XML
+
+        .. example :: self.parse_features("n-p---na-")
+
+        :param features: A string containing morphological informations
+        :type features: str
+        :return: Parsed features
+        :rtype: dict
+        """
+        raise NotImplementedError("Parse Features has not been implemented for this class")

pyperseus_treebank/latin.py

@@ -0,0 +1,66 @@
+"""
+
+..source :: https://github.com/perseids-project/perseids_treebanking/blob/ae0305138dacc4a89c5fe6b0f086a4b3b1efdc92/transformations/aldt-util.xsl
+"""
+from .base import Token
+
+
+_NUMBER = {"s": "Sing", "p": "Plur"}
+_TENSE = {"p": "Pres", "f": "Fut", "r": "Perf", "l": "PQP", "i": "Imp"}
+_MOOD = {"i": "Ind", "s": "Sub", "m": "Imp", "g": "Ger", "p": "Part", "u": "Sup", "n": "Inf"}
+_VOICE = {"a": "Act", "p": "Pass", "d": "Dep"}
+_GENDER = {"f": "Fem", "m": "Masc", "n": "Neut", "c": "Com"}
+_CASE = {"g": "Gen", "d": "Dat", "a": "Acc", "v": "Voc", "n": "Nom", "b": "Abl", "i": "Ins", "l": "Loc"}
+_DEGREE = {"p": "Pos", "c": "Comp", "s": "Sup"}
+
+
+class LatinToken(Token):
+
+    def parse_features(self, features):
+        """ Parse features from the POSTAG of Perseus Latin XML
+
+        .. example :: self.parse_features("n-p---na-")
+
+        :param features: A string containing morphological informations
+        :type features: str
+        :return: Parsed features
+        :rtype: dict
+        """
+
+        feats = {}
+
+        self.pos = features[0]
+
+        # Person handling : 3 possibilities
+        if features[1] != "-":
+            feats["Person"] = features[1]
+
+        # Number handling : two possibilities
+        if features[2] != "-":
+            feats["Number"] = _NUMBER[features[2]]
+
+        # Tense
+        if features[3] != "-":
+            feats["Tense"] = _TENSE[features[3]]
+
+        # Mood
+        if features[4] != "-":
+            feats["Mood"] = _MOOD[features[4]]
+
+        # Voice
+        if features[5] != "-":
+            feats["Voice"] = _VOICE[features[5]]
+
+        # Tense
+        if features[6] != "-":
+            feats["Gender"] = _GENDER[features[6]]
+
+        # Tense
+        if features[7] != "-":
+            feats["Case"] = _CASE[features[7]]
+
+        # Degree
+        if features[8] != "-":
+            feats["Degree"] = _DEGREE[features[8]]
+
+        return feats

requirements.txt

@@ -0,0 +1 @@
+lxml

setup.py

@@ -0,0 +1,21 @@
+from setuptools import setup, find_packages
+
+version = "0.0.1"
+
+setup(
+    name='pyperseus_treebank',
+    version=version,
+    packages=find_packages(exclude=["tests"]),
+    url='https://github.com/ponteineptique/pyperseus-treebank',
+    license='GNU GPL v2',
+    author='Thibault Clerice',
+    author_email='leponteineptique@gmail.com',
+    description='Perseus XML Treebank Parser',
+    test_suite="tests",
+    install_requires=[
+        "lxml"
+    ],
+    test_requires=[
+        "coverage==4.4.1"
+    ]
+)

tests/test_latin.py

@@ -0,0 +1,20 @@
+from pyperseus_treebank.latin import LatinToken
+from unittest import TestCase
+
+
+class TestLatinToken(TestCase):
+    def test_features(self):
+        """
+
+        Source : <word id="4" form="cano" lemma="cano" postag="v1spia---" relation="PRED" head="0"/>
+        :return:
+        """
+        cano = LatinToken(4, "cano", "cano", 0, "v1spia---", rel="HEAD")
+        self.assertEqual(cano.pos, "v")
+        self.assertEqual(cano.features, {
+            "Person": "1",
+            "Number": "Sing",
+            "Tense": "Pres",
+            "Mood": "Ind",
+            "Voice": "Act"
+        })