pythainlp 5.0.0.dev1__tar.gz → 5.0.0.dev2__tar.gz

{pythainlp-5.0.0.dev1/pythainlp.egg-info → pythainlp-5.0.0.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: pythainlp
-Version: 5.0.0.dev1
+Version: 5.0.0.dev2
 Summary: Thai Natural Language Processing library
 Home-page: https://github.com/PyThaiNLP/pythainlp
 Author: PyThaiNLP

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/README.md

@@ -29,7 +29,7 @@ PyThaiNLP เป็นไลบารีภาษาไพทอนสำหร
 
 ## Getting Started
 
-- PyThaiNLP 2 requires Python 3.6+. Python 2.7 users can use PyThaiNLP 1.6. See [2.0 change log](https://github.com/PyThaiNLP/pythainlp/issues/118) | [Upgrading from 1.7](https://pythainlp.github.io/docs/2.0/notes/pythainlp-1_7-2_0.html) | [Upgrading ThaiNER from 1.7](https://github.com/PyThaiNLP/pythainlp/wiki/Upgrade-ThaiNER-from-PyThaiNLP-1.7-to-PyThaiNLP-2.0)
+- PyThaiNLP 2 requires Python 3.7+. Python 2.7 users can use PyThaiNLP 1.6. See [2.0 change log](https://github.com/PyThaiNLP/pythainlp/issues/118) | [Upgrading from 1.7](https://pythainlp.github.io/docs/2.0/notes/pythainlp-1_7-2_0.html) | [Upgrading ThaiNER from 1.7](https://github.com/PyThaiNLP/pythainlp/wiki/Upgrade-ThaiNER-from-PyThaiNLP-1.7-to-PyThaiNLP-2.0)
 - [PyThaiNLP Get Started notebook](https://pythainlp.github.io/tutorials/notebooks/pythainlp_get_started.html) | [API document](https://pythainlp.github.io/docs) | [Tutorials](https://pythainlp.github.io/tutorials)
 - [Official website](https://pythainlp.github.io/) | [PyPI](https://pypi.org/project/pythainlp/) | [Facebook page](https://www.facebook.com/pythainlp/)
 - [Who uses PyThaiNLP?](https://github.com/PyThaiNLP/pythainlp/blob/dev/INTHEWILD.md)
@@ -148,13 +148,18 @@ or by BibTeX entry:
 
 ``` bib
 @misc{pythainlp,
-    author       = {Wannaphong Phatthiyaphaibun and Korakot Chaovavanich and Charin Polpanumas and Arthit Suriyawongkul and Lalita Lowphansirikul and Pattarawat Chormai},
-    title        = {{PyThaiNLP: Thai Natural Language Processing in Python}},
-    month        = Jun,
-    year         = 2016,
-    doi          = {10.5281/zenodo.3519354},
-    publisher    = {Zenodo},
-    url          = {http://doi.org/10.5281/zenodo.3519354}
+    title = "{P}y{T}hai{NLP}: {T}hai Natural Language Processing in {P}ython",
+    author = "Phatthiyaphaibun, Wannaphong  and
+      Chaovavanich, Korakot  and
+      Polpanumas, Charin  and
+      Suriyawongkul, Arthit  and
+      Lowphansirikul, Lalita  and
+      Chormai, Pattarawat",
+    month = jun,
+    year = "2016",
+    doi = {10.5281/zenodo.3519354},
+    publisher = {Zenodo},
+    url = {http://doi.org/10.5281/zenodo.3519354}
 }
 ```
 
@@ -166,7 +171,7 @@ and its BibTeX entry:
 
 ```bib
 @inproceedings{phatthiyaphaibun-etal-2023-pythainlp,
-    title = "{P}y{T}hai{NLP}: {T}hai Natural Language Processing in Python",
+    title = "{P}y{T}hai{NLP}: {T}hai Natural Language Processing in {P}ython",
     author = "Phatthiyaphaibun, Wannaphong  and
       Chaovavanich, Korakot  and
       Polpanumas, Charin  and

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/__init__.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
-__version__ = "5.0.0dev1"
+__version__ = "5.0.0dev2"
 
 thai_consonants = "กขฃคฅฆงจฉชซฌญฎฏฐฑฒณดตถทธนบปผฝพฟภมยรลวศษสหฬอฮ"  # 44 chars
 

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/__main__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 import argparse
 import sys

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/ancient/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Ancient versions of the Thai language

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/ancient/aksonhan.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from pythainlp.util import Trie
 from pythainlp import thai_consonants, thai_tonemarks

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai text augment

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/lm/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Language Models

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/lm/fasttext.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 import itertools
 from typing import List, Tuple

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/lm/phayathaibert.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 from typing import List

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/lm/wangchanberta.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 from typing import List

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/word2vec/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Word2Vec

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/word2vec/bpemb_wv.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List, Tuple
 from pythainlp.augment.word2vec.core import Word2VecAug

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/word2vec/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List, Tuple
 import itertools

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/word2vec/ltw2v.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List, Tuple
 from pythainlp.augment.word2vec.core import Word2VecAug

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/word2vec/thai2fit.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List, Tuple
 from pythainlp.augment.word2vec.core import Word2VecAug

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/augment/wordnet.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thank https://dev.to/ton_ami/text-data-augmentation-synonym-replacement-4h8l

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/benchmarks/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Performance benchmarking.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/benchmarks/word_tokenization.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 import re

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/chat/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 pythainlp.chat

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/chat/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 import torch
 

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/classify/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 pythainlp.classify

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/classify/param_free.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 import gzip

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cli/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """Command line helpers."""
 import sys

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cli/benchmark.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 import argparse

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cli/data.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Command line for PyThaiNLP's dataset/corpus management.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cli/soundex.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Command line for PyThaiNLP's soundex.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cli/tag.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Command line for PyThaiNLP's taggers.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cli/tokenize.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Command line for PyThaiNLP's tokenizers.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/cls/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 pythainlp.cls

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/coref/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 PyThaiNLP Coreference Resolution

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/coref/_fastcoref.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List
 import spacy

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/coref/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List
 

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/coref/han_coref.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 import spacy
 from pythainlp.coref._fastcoref import FastCoref

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Corpus related functions.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/common.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 """

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/conceptnet.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Get data from ConceptNet API at http://conceptnet.io

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Corpus related functions.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/icu.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Provides an optional word list from International Components for Unicode (ICU) dictionary.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/oscar.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai unigram word frequency from OSCAR Corpus (words tokenized using ICU)

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/th_en_translit.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai-English Transliteration Dictionary v1.4

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/tnc.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project.
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project.
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai National Corpus word frequency

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/ttc.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai Textbook Corpus (TTC) word frequency

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/util.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Tool for creating word lists

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/volubilis.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Provides an optional word list from the Volubilis dictionary.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/wikipedia.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Provides an optional word list from Thai Wikipedia titles.

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/corpus/wordnet.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 NLTK WordNet wrapper

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/el/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 pythainlp.el

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/el/_multiel.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/el/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 from typing import List, Union
 

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/generate/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai Text Generation

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/generate/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Text generator using n-gram language model

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/generate/thai2fit.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Thai2fit: Thai Wikipeida Language Model for Text Generation

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/generate/wangchanglm.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 import re
 import torch

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/khavee/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 __all__ = ["KhaveeVerifier"]

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/khavee/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 # ruff: noqa: C901
 

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/khavee/example.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 import core

pythainlp-5.0.0.dev2/pythainlp/morpheme/__init__.py

@@ -0,0 +1,13 @@
+# -*- coding: utf-8 -*-
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
+# SPDX-License-Identifier: Apache-2.0
+
+"""
+PyThaiNLP morpheme
+"""
+__all__ = [
+    "nighit",
+    "is_native_thai"
+]
+from pythainlp.morpheme.word_formation import nighit
+from pythainlp.morpheme.thaiwordcheck import is_native_thai

{pythainlp-5.0.0.dev1/pythainlp/util → pythainlp-5.0.0.dev2/pythainlp/morpheme}/thaiwordcheck.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 """
 Check if a word is a "native Thai word"
@@ -8,7 +8,8 @@ Adapted from
 https://github.com/wannaphong/open-thai-nlp-document/blob/master/check_thai_word.md
 
 References
-- ทีมงานทรูปลูกปัญญา 2015. ลักษณะของคำไทยแท้ http://www.trueplookpanya.com/learning/detail/30589-043067
+- ทีมงานทรูปลูกปัญญา 2015. ลักษณะของคำไทยแท้ \
+    http://www.trueplookpanya.com/learning/detail/30589-043067
 - วารุณี บำรุงรส 2010. คำไทยแท้ https://www.gotoknow.org/posts/377619
 """
 import re

pythainlp-5.0.0.dev2/pythainlp/morpheme/word_formation.py

@@ -0,0 +1,56 @@
+# -*- coding: utf-8 -*-
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
+# SPDX-License-Identifier: Apache-2.0
+from pythainlp import thai_consonants
+
+
+def nighit(w1: str, w2: str) -> str:
+    """
+    Nighit (นิคหิต or  ํ ) is the niggahita in Thai language for create new \
+    words from Pali language in Thai.
+    The function use simple method to create new Thai word from two words \
+    that the root is from Pali language.
+
+    Read more: https://www.trueplookpanya.com/learning/detail/1180
+
+    :param str w1: A Thai word that has a nighit.
+    :param str w2: A Thai word.
+    :return: Thai word.
+    :rtype: str
+    :Example:
+    ::
+        from pythainlp.morpheme import nighit
+
+        assert nighit("สํ","คีต")=="สังคีต"
+        assert nighit("สํ","จร")=="สัญจร"
+        assert nighit("สํ","ฐาน")=="สัณฐาน"
+        assert nighit("สํ","นิษฐาน")=="สันนิษฐาน"
+        assert nighit("สํ","ปทา")=="สัมปทา"
+        assert nighit("สํ","โยค")=="สังโยค"
+    """
+    if not str(w1).endswith('ํ') and len(w1) != 2:
+        raise NotImplementedError(f"The function doesn't support {w1}.")
+    list_w1 = list(w1)
+    list_w2 = list(w2)
+    newword = list()
+    newword.append(list_w1[0])
+    newword.append("ั")
+    consonant_start = [i for i in list_w2 if i in set(thai_consonants)][0]
+    if consonant_start in ["ก", "ช", "ค", "ข", "ง"]:
+        newword.append("ง")
+    elif consonant_start in ["จ", "ฉ", "ช", "ฌ"]:
+        newword.append("ญ")
+    elif consonant_start in ["ฎ", "ฐ", "ฑ", "ณ"]:
+        newword.append("ณ")
+    elif consonant_start in ["ด", "ถ", "ท", "ธ", "น"]:
+        newword.append("น")
+    elif consonant_start in ["ป", "ผ", "พ", "ภ"]:
+        newword.append("ม")
+    elif consonant_start in ["ย", "ร", "ล", "ฬ", "ว", "ศ", "ษ", "ส", "ห"]:
+        newword.append("ง")
+    else:
+        raise NotImplementedError(f"""
+        The function doesn't support {w1} and {w2}.
+        """)
+    newword.extend(list_w2)
+    return ''.join(newword)

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/parse/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 """

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/parse/core.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 from typing import List, Union

{pythainlp-5.0.0.dev1 → pythainlp-5.0.0.dev2}/pythainlp/phayathaibert/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# SPDX-FileCopyrightText: Copyright 2016-2023 PyThaiNLP Project
+# SPDX-FileCopyrightText: Copyright 2016-2024 PyThaiNLP Project
 # SPDX-License-Identifier: Apache-2.0
 
 """