Merge pull request #34 from SekouDiaoNlp/dev

Implemented complete CLI with ability to save the Lexical Entries to a json file.

pylexique/cli.py

@@ -4,17 +4,17 @@
 import sys
 import click
 import json
-import joblib
 import logging
 from pylexique import Lexique383, LexItem
 from pprint import pprint
+from collections import defaultdict
 
 
 @click.command(context_settings=dict(help_option_names=["-h", "--help"]))
 @click.argument('words', nargs=-1)
 @click.option('-o', '--output',
               default=None,
-              help=("Path of the filename for storing the lexical entries"),
+              help="Path of the json filename for storing the lexical entries",
               type=click.STRING)
 def main(words, output):
     """Pylexique is a Python wrapper around Lexique83.
@@ -36,23 +36,22 @@ def main(words, output):
     logger.setLevel(logging.INFO)
 
     LEXIQUE = Lexique383()
-    results = {}
+    results = defaultdict(list)
     for word in words:
-        results[word] = LEXIQUE.lexique[word]
+        results[word].append(LEXIQUE.lexique[word])
+
+        for i, element in enumerate(results[word]):
+            if isinstance(element, LexItem):
+                results[word][i] = element.to_dict()
+                continue
+            for index, item in enumerate(element):
+                results[word][i][index] = item.to_dict()
     if output:
         with open(output, 'w', encoding='utf-8') as file:
             json.dump(results, file, indent=4)
-            pprint('The Lexical Items have been successfully saved to {0} by pylexique.cli.main.'.format(output))
+            print('The Lexical Items have been successfully saved to {0} by pylexique.cli.main.'.format(output))
     else:
-        if isinstance(words, tuple):
-            for word in words:
-                if isinstance(results[word], list):
-                    for elmt in results[word]:
-                        pprint(elmt.to_dict())
-                        print('\n\n')
-                else:
-                    pprint(results[word].to_dict())
-                    print('\n\n')
+        print(json.dumps(results, indent=4))
     return
 
 

pylexique/pylexique.py

@@ -4,11 +4,10 @@
 
 from collections import OrderedDict, defaultdict
 import pkg_resources
-
+# import faster_than_csv as csv
 from dataclasses import dataclass
 from typing import ClassVar
 from time import time
-from .utils import set_save_folder
 
 _RESOURCE_PACKAGE = __name__
 
@@ -161,9 +160,6 @@ def _convert_entries(row_fields):
                     try:
                         value = int(value)
                     except ValueError:
-                        print(
-                            "the value {} is of the wrong type for the attribute '{}'. Keeping value as string.\n".format(
-                                value, attr))
                         errors[row_fields[0]].append({attr: value})
                         value = value
             converted_row_fields.append(value)

pylexique/utils.py

@@ -1,49 +1,4 @@
 from dataclasses import dataclass
-from pprint import pprint
-from inspect import getmembers
-from types import FunctionType
-
-
-def set_save_folder(folder):
-    """
-    Sets the folder in which the LexicalItems will be downloaded and saved.
-
-    :param folder: string.
-        Folder path.
-    :return: string.
-        Folder path.
-    """
-    if not folder:
-        folder = os.path.join(os.path.expanduser("~"), 'Documents', 'pylexique')
-    else:
-        folder = os.path.join(folder, 'pylexique')
-    return folder
-
-def attributes(obj):
-    disallowed_names = {
-      name for name, value in getmembers(type(obj))
-        if isinstance(value, FunctionType)}
-    return {
-      name: getattr(obj, name) for name in dir(obj)
-        if name[0:1] != '_' and name not in disallowed_names and hasattr(obj, name)}
-
-def show_attributes(object):
-    temp = vars(object)
-    for item in temp:
-        print(item, ':', temp[item])
-
-def print_attributes(obj):
-    pprint(attributes(obj))
-
-
-def vdir(obj):
-    """
-    | This function pretty-display the lexical items in Lexique383.
-
-    :param obj:
-    :return:
-    """
-    return [x for x in dir(obj) if not x.startswith('__')]
 
 
 def dataclass_with_default_init(_cls=None, *args, **kwargs):

requirements.txt

@@ -1,7 +1,6 @@
 colorama
 tqdm
+# faster_than_csv
 Click>=7.1
-pandas
 setuptools~=49.6.0
-pytest~=6.2.3
 joblib~=1.0.1

setup.py

@@ -11,7 +11,7 @@
 with open('HISTORY.rst') as history_file:
     history = history_file.read()
 
-requirements = ['joblib', 'colorama', 'tqdm', 'Click>=7.1', 'joblib', 'pandas']
+requirements = ['joblib', 'colorama', 'tqdm', 'Click>=7.1'] # integrate 'faster_than_csv'
 
 setup_requirements = ['pytest-runner', ]