Number converters: convert more than one digit at a time

pythainlp/number/__init__.py

@@ -3,10 +3,23 @@
 Number conversions between Thai digits, Arabic digits, and Thai words
 """
 
-from .thainum import (
-    bahttext,
-    num_to_thaiword,
+from .digitconv import (
+    arabic_digit_to_thai_digit,
+    digit_to_text,
+    text_to_arabic_digit,
+    text_to_thai_digit,
+    thai_digit_to_arabic_digit,
 )
+from .numtoword import bahttext, num_to_thaiword
 from .wordtonum import thaiword_to_num
 
-__all__ = ["bahttext", "num_to_thaiword", "thaiword_to_num"]
+__all__ = [
+    "bahttext",
+    "num_to_thaiword",
+    "thaiword_to_num",
+    "arabic_digit_to_thai_digit",
+    "digit_to_text",
+    "text_to_arabic_digit",
+    "text_to_thai_digit",
+    "thai_digit_to_arabic_digit",
+]

pythainlp/number/digitconv.py

@@ -0,0 +1,132 @@
+# -*- coding: utf-8 -*-
+"""
+Convert digits
+"""
+
+_arabic_thai = {
+    "0": "๐",
+    "1": "๑",
+    "2": "๒",
+    "3": "๓",
+    "4": "๔",
+    "5": "๕",
+    "6": "๖",
+    "7": "๗",
+    "8": "๘",
+    "9": "๙",
+}
+
+_thai_arabic = {
+    "๐": "0",
+    "๑": "1",
+    "๒": "2",
+    "๓": "3",
+    "๔": "4",
+    "๕": "5",
+    "๖": "6",
+    "๗": "7",
+    "๘": "8",
+    "๙": "9",
+}
+
+_digit_spell = {
+    "0": "ศูนย์",
+    "1": "หนึ่ง",
+    "2": "สอง",
+    "3": "สาม",
+    "4": "สี่",
+    "5": "ห้า",
+    "6": "หก",
+    "7": "เจ็ด",
+    "8": "แปด",
+    "9": "เก้า",
+}
+
+_spell_digit = {
+    "ศูนย์": "0",
+    "หนึ่ง": "1",
+    "สอง": "2",
+    "สาม": "3",
+    "สี่": "4",
+    "ห้า": "5",
+    "หก": "6",
+    "เจ็ด": "7",
+    "แปด": "8",
+    "เก้า": "9",
+}
+
+
+def thai_digit_to_arabic_digit(text):
+    """
+    :param str text: Text with Thai digits such as '๑', '๒', '๓'
+    :return: Text with Thai digits being converted to Arabic digits such as '1', '2', '3'
+    """
+    if not text:
+        return ""
+
+    newtext = []
+    for ch in text:
+        if ch in _thai_arabic:
+            newtext.append(_thai_arabic[ch])
+        else:
+            newtext.append(ch)
+
+    return "".join(newtext)
+
+
+def arabic_digit_to_thai_digit(text):
+    """
+    :param str text: Text with Arabic digits such as '1', '2', '3'
+    :return: Text with Arabic digits being converted to Thai digits such as '๑', '๒', '๓'
+    """
+    if not text:
+        return ""
+
+    newtext = []
+    for ch in text:
+        if ch in _arabic_thai:
+            newtext.append(_arabic_thai[ch])
+        else:
+            newtext.append(ch)
+
+    return "".join(newtext)
+
+
+def digit_to_text(text):
+    """
+    :param str text: Text with digits such as '1', '2', '๓', '๔'
+    :return: Text with digits being spelled out in Thai
+    """
+    if not text:
+        return ""
+
+    newtext = []
+    for ch in text:
+        if ch in _thai_arabic:
+            ch = _thai_arabic[ch]
+
+        if ch in _digit_spell:
+            newtext.append(_digit_spell[ch])
+        else:
+            newtext.append(ch)
+
+    return "".join(newtext)
+
+
+def text_to_arabic_digit(text):
+    """
+    :param text: A digit spelled out in Thai
+    :return: An Arabic digit such as '1', '2', '3'
+    """
+    if not text or text not in _spell_digit:
+        return ""
+
+    return _spell_digit[text]
+
+
+def text_to_thai_digit(text):
+    """
+    :param text: A digit spelled out in Thai
+    :return: A Thai digit such as '๑', '๒', '๓'
+    """
+    return arabic_digit_to_thai_digit(text_to_arabic_digit(text))

pythainlp/number/thainum.py → pythainlp/number/numtoword.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 """
-Convert Thai numbers
+Convert number value to Thai read out
 
 Adapted from
 http://justmindthought.blogspot.com/2012/12/code-php.html
@@ -9,74 +9,6 @@
 
 __all__ = ["bahttext", "num_to_thaiword"]
 
-_p = [
-    ["ภาษาไทย", "ตัวเลข", "เลขไทย"],
-    ["หนึ่ง", "1", "๑"],
-    ["สอง", "2", "๒"],
-    ["สาม", "3", "๓"],
-    ["สี่", "4", "๔"],
-    ["ห้า", "5", "๕"],
-    ["หก", "6", "๖"],
-    ["หก", "7", "๗"],
-    ["แปด", "8", "๘"],
-    ["เก้า", "9", "๙"],
-]
-
-
-# เลขไทยสู่เลขอารบิก
-def thai_num_to_num(text):
-    """
-    :param str text: Thai number characters such as '๑', '๒', '๓'
-    :return: universal numbers such as '1', '2', '3'
-    """
-    thaitonum = dict((x[2], x[1]) for x in _p[1:])
-    return thaitonum[text]
-
-
-def thai_num_to_text(text):
-    """
-    :param str text: Thai number characters such as '๑', '๒', '๓'
-    :return: Thai numbers, spelled out in Thai
-    """
-    thaitonum = dict((x[2], x[0]) for x in _p[1:])
-    return thaitonum[text]
-
-
-def num_to_thai_num(text):
-    """
-    :param text: universal numbers such as '1', '2', '3'
-    :return: Thai number characters such as '๑', '๒', '๓'
-    """
-    thaitonum = dict((x[1], x[2]) for x in _p[1:])
-    return thaitonum[text]
-
-
-def num_to_text(text):
-    """
-    :param text: universal numbers such as '1', '2', '3'
-    :return: Thai numbers, spelled out in Thai
-    """
-    thaitonum = dict((x[1], x[0]) for x in _p[1:])
-    return thaitonum[text]
-
-
-def text_to_num(text):
-    """
-    :param text: Thai numbers, spelled out in Thai
-    :return: universal numbers such as '1', '2', '3'
-    """
-    thaitonum = dict((x[0], x[1]) for x in _p[1:])
-    return thaitonum[text]
-
-
-def text_to_thai_num(text):
-    """
-    :param text: Thai numbers, spelled out in Thai
-    :return: Thai numbers such as '๑', '๒', '๓'
-    """
-    thaitonum = dict((x[0], x[2]) for x in _p[1:])
-    return thaitonum[text]
-
 
 def bahttext(number):
     """

pythainlp/number/wordtonum.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 """
-Convert number in words to a computablenumber value
+Convert number in words to a computable number value
 
 Adapted from Korakot Chaovavanich's notebook
 https://colab.research.google.com/drive/148WNIeclf0kOU6QxKd6pcfwpSs8l-VKD#scrollTo=EuVDd0nNuI8Q

tests/__init__.py

@@ -21,7 +21,16 @@
 from pythainlp.keywords import find_keyword
 from pythainlp.ner import ThaiNameRecognizer
 from pythainlp.ner.locations import tag_provinces
-from pythainlp.number import bahttext, num_to_thaiword, thaiword_to_num
+from pythainlp.number import (
+    arabic_digit_to_thai_digit,
+    bahttext,
+    digit_to_text,
+    num_to_thaiword,
+    text_to_arabic_digit,
+    text_to_thai_digit,
+    thai_digit_to_arabic_digit,
+    thaiword_to_num,
+)
 from pythainlp.rank import rank
 from pythainlp.romanization import romanize
 from pythainlp.sentiment import sentiment
@@ -172,6 +181,26 @@ def test_number(self):
         self.assertEqual(thaiword_to_num(""), None)
         self.assertEqual(thaiword_to_num(None), None)
 
+        self.assertEqual(arabic_digit_to_thai_digit("ไทยแลนด์ 4.0"), "ไทยแลนด์ ๔.๐")
+        self.assertEqual(arabic_digit_to_thai_digit(""), "")
+        self.assertEqual(arabic_digit_to_thai_digit(None), "")
+
+        self.assertEqual(thai_digit_to_arabic_digit("๔๐๔ Not Found"), "404 Not Found")
+        self.assertEqual(thai_digit_to_arabic_digit(""), "")
+        self.assertEqual(thai_digit_to_arabic_digit(None), "")
+
+        self.assertEqual(digit_to_text("RFC 7258"), "RFC เจ็ดสองห้าแปด")
+        self.assertEqual(digit_to_text(""), "")
+        self.assertEqual(digit_to_text(None), "")
+
+        self.assertEqual(text_to_arabic_digit("เจ็ด"), "7")
+        self.assertEqual(text_to_arabic_digit(""), "")
+        self.assertEqual(text_to_arabic_digit(None), "")
+
+        self.assertEqual(text_to_thai_digit("เก้า"), "๙")
+        self.assertEqual(text_to_thai_digit(""), "")
+        self.assertEqual(text_to_thai_digit(None), "")
+
     # ### pythainlp.rank
 
     def test_rank(self):