interscript 0.1.2 → 0.1.7

data/maps/royin-tha-Thai-Latn-1968.yaml

@@ -0,0 +1,183 @@
+---
+authority_id: royin
+id: 1968-chained
+language: iso-639-2:tha
+source_script: Thai
+destination_script: Latn
+name: Royal Thai General System of Transcription (1968)
+alias:
+  ogc11122:
+    code: tha_Thai2Latn_RIT_1968
+    description: Thai Royal Institute of Thailand 1968 System (out of date)
+url: http://www.royin.go.th/wp-content/uploads/royin-ebook/276/FileUpload/758_6484.pdf
+creation_date: 1968
+adoption_date:
+description: |
+  This map loads two external maps to convert Thai text first into phonemic Thai,
+  and then into IPA transcription.
+
+  The IPA transcription will then be handled by this map, and converted into
+  Royal Thai General System of Transcription (1968).
+
+  The first two parts are done via two external maps.
+
+notes: |
+  The conversion from Thai to Phonemic Thai is still work-in-progress.
+
+tests:
+  - source: "สะพาน"
+    expected: "saphan"
+  - source: "ลานตา"
+    expected: "lanta"
+  - source: "บาง"
+    expected: "bang"
+  - source: "สมุทร"
+    expected: "samut"
+  - source: "ลำ"
+    expected: "lam"
+  - source: "สิงห์"
+    expected: "sing"
+  - source: "บุรี"
+    expected: "buri"
+  - source: "สตึก"
+    expected: "satuk"
+  - source: "พืช"
+    expected: "phut"
+  - source: "บรบือ"
+    expected: "borabu"
+  - source: "ภู"
+    expected: "phu"
+  - source: "ปะนาเระ"
+    expected: "panare"
+  - source: "เพ็ญ"
+    expected: "phen"
+  # - source: "เขน"
+  #   expected: "khen"
+  - source: "แซะ"
+    expected: "sae"
+  # - source: "สะแก"
+  #   expected: "sakae"
+  - source: "พะโต๊ะ"
+    expected: "phato"
+  - source: "ลพ"
+    expected: "lop"
+  # - source: "สามโก้"
+  #   expected: "samko"
+  - source: "เกาะ"
+    expected: "ko"
+  - source: "บ่อ"
+    expected: "bo"
+  - source: "เซอะ"
+    expected: "soe"
+  - source: "อำเภอ"
+    expected: "amphoe"
+  - source: "เนิน"
+    expected: "noen"
+  # - source: "เพียะ"
+  #   expected: "phia"
+  - source: "เทียน"
+    expected: "thian"
+  # - source: "เกือะ"
+  #   expected: "kua"
+  - source: "เมือง"
+    expected: "muang"
+  # - source: "ผัวะ"
+  #   expected: "phua"
+  - source: "บัว"
+    expected: "bua"
+  # - source: "ควน"
+  #   expected: "khuan"
+  - source: "ใหญ่"
+    expected: "yai"
+  # - source: "ไผ่"
+  #   expected: "phai"
+  - source: "ชัย"
+    expected: "chai"
+  - source: "ไทย"
+    expected: "thai"
+  # - source: "ปาย"
+  #   expected: "pai"
+  - source: "เจ้า"
+    expected: "chao"
+  - source: "ข้าว"
+    expected: "khao"
+  # - source: "กุย"
+  #   expected: "kui"
+  - source: "โดย"
+    expected: "doi"
+  # - source: "ดอย"
+  #   expected: "doi"
+  # - source: "งิ้ว"
+  #   expected: "ngiu"
+  - source: "เร็ว"
+    expected: "reo"
+  # - source: "เลว"
+  #   expected: "leo"
+  # - source: "เลย"
+  #   expected: "loei"
+  # - source: "เดือย"
+  #   expected: "duai"
+  # - source: "ห้วย"
+  #   expected: "huai"
+  - source: "แมว"
+    expected: "maeo"
+  - source: "เขียว"
+    expected: "khieu"
+
+chain: ["var-tha-Thai-Thai-phonemic" ,"var-tha-Thai-Zsym-ipa"]
+
+map:
+  title-case: false
+  word_separator: " "
+
+  rules:
+    - pattern: '[˩˨˧˦˥]'
+      result : ''
+    - pattern: '^'
+      result: '.'
+
+  postrules:
+    - pattern: '\.'
+      result:  ''
+
+  characters:
+
+  dictionary:
+
+    '̯': ''
+    '̚': ''
+
+    'ʔ': ''
+    'ː': ''
+
+    't͡ɕʰ': 'ch'
+    't͡ɕ': 'ch'
+    'ŋ': 'ng'
+    'j': 'y'
+    'ɔ': 'o'
+    'ɤ': 'oe'
+    'ɛ': 'ae'
+    'ɯ': 'u'
+    'ʰ': 'h'
+
+    'aːw': 'ao'
+    'aw': 'ao'
+    'a̯w': 'eu'
+    'eːw': 'eo'
+    'ew': 'eo'
+    'ɛːw': 'aeo'
+    'ɛw': 'aeo'
+    'iːw': 'iu'
+    'iw': 'iu'
+
+    'aːj': 'ai'
+    'aj': 'ai'
+    'a̯j': 'ai'
+    'ɔːj': 'oi'
+    'ɔj': 'oi'
+    'oːj': 'oi'
+    'oj': 'oi'
+    'ɤːj': 'oei'
+    'ɤj': 'oei'
+    'uːj': 'ui'
+    'uj': 'ui'

data/maps/royin-tha-Thai-Latn-1999-chained.yaml

@@ -0,0 +1,180 @@
+---
+authority_id: royin
+id: 1999-chained
+language: iso-639-2:tha
+source_script: Thai
+destination_script: Latn
+name: Royal Thai General System of Transcription (1999)
+url: http://www.royin.go.th/wp-content/uploads/royin-ebook/276/FileUpload/758_6484.pdf
+creation_date: 1999
+adoption_date:
+description: |
+  This map loads two external maps to convert Thai text first into phonemic Thai,
+  and then into IPA transcription.
+
+  The IPA transcription will then be handled by this map, and converted into
+  Royal Thai General System of Transcription (1999)
+
+  The first two parts are done via two external maps.
+
+notes: |
+  The conversion from Thai to Phonemic Thai is still work-in-progress.
+
+tests:
+  - source: "สะพาน"
+    expected: "saphan"
+  - source: "ลานตา"
+    expected: "lanta"
+  - source: "บาง"
+    expected: "bang"
+  - source: "สมุทร"
+    expected: "samut"
+  - source: "ลำ"
+    expected: "lam"
+  - source: "สิงห์"
+    expected: "sing"
+  - source: "บุรี"
+    expected: "buri"
+  - source: "สตึก"
+    expected: "satuek"
+  - source: "พืช"
+    expected: "phuet"
+  - source: "บรบือ"
+    expected: "borabue"
+  - source: "ภู"
+    expected: "phu"
+  - source: "ปะนาเระ"
+    expected: "panare"
+  - source: "เพ็ญ"
+    expected: "phen"
+  # - source: "เขน"
+  #   expected: "khen"
+  - source: "แซะ"
+    expected: "sae"
+  # - source: "สะแก"
+  #   expected: "sakae"
+  - source: "พะโต๊ะ"
+    expected: "phato"
+  - source: "ลพ"
+    expected: "lop"
+  # - source: "สามโก้"
+  #   expected: "samko"
+  - source: "เกาะ"
+    expected: "ko"
+  - source: "บ่อ"
+    expected: "bo"
+  - source: "เซอะ"
+    expected: "soe"
+  - source: "อำเภอ"
+    expected: "amphoe"
+  - source: "เนิน"
+    expected: "noen"
+  # - source: "เพียะ"
+  #   expected: "phia"
+  - source: "เทียน"
+    expected: "thian"
+  # - source: "เกือะ"
+  #   expected: "kuea"
+  - source: "เมือง"
+    expected: "mueang"
+  # - source: "ผัวะ"
+  #   expected: "phua"
+  - source: "บัว"
+    expected: "bua"
+  # - source: "ควน"
+  #   expected: "khuan"
+  - source: "ใหญ่"
+    expected: "yai"
+  # - source: "ไผ่"
+  #   expected: "phai"
+  - source: "ชัย"
+    expected: "chai"
+  - source: "ไทย"
+    expected: "thai"
+  # - source: "ปาย"
+  #   expected: "pai"
+  - source: "เจ้า"
+    expected: "chao"
+  - source: "ข้าว"
+    expected: "khao"
+  # - source: "กุย"
+  #   expected: "kui"
+  - source: "โดย"
+    expected: "doi"
+  # - source: "ดอย"
+  #   expected: "doi"
+  # - source: "งิ้ว"
+  #   expected: "ngio"
+  - source: "เร็ว"
+    expected: "reo"
+  # - source: "เลว"
+  #   expected: "leo"
+  # - source: "เลย"
+  #   expected: "loei"
+  # - source: "เดือย"
+  #   expected: "dueai"
+  # - source: "ห้วย"
+  #   expected: "huai"
+  - source: "แมว"
+    expected: "maeo"
+  - source: "เขียว"
+    expected: "khiao"
+
+
+chain: ["var-tha-Thai-Thai-phonemic" ,"var-tha-Thai-Zsym-ipa"]
+
+map:
+  title-case: false
+  word_separator: " "
+
+  rules:
+    - pattern: '[˩˨˧˦˥]'
+      result : ''
+    - pattern: '^'
+      result: '.'
+
+  postrules:
+    - pattern: '\.'
+      result:  ''
+
+  characters:
+
+  dictionary:
+
+    '̯': ''
+    '̚': ''
+
+    'ʔ': ''
+    'ː': ''
+
+    't͡ɕʰ': 'ch'
+    't͡ɕ': 'ch'
+    'ŋ': 'ng'
+    'j': 'y'
+    'ɔ': 'o'
+    'ɤ': 'oe'
+    'ɛ': 'ae'
+    'ɯ': 'ue' # New spelling, was u in 1968
+    'ʰ': 'h'
+
+    'aːw': 'ao'
+    'aw': 'ao'
+    'a̯w': 'ao' # New spelling, was eu in 1968
+    'eːw': 'eo'
+    'ew': 'eo'
+    'ɛːw': 'aeo'
+    'ɛw': 'aeo'
+    'iːw': 'io' # New spelling, was iu in 1968
+    'iw': 'io' # New spelling, was iu in 1968
+
+    'aːj': 'ai'
+    'aj': 'ai'
+    'a̯j': 'ai'
+    'ɔːj': 'oi'
+    'ɔj': 'oi'
+    'oːj': 'oi'
+    'oj': 'oi'
+    'ɤːj': 'oei'
+    'ɤj': 'oei'
+    'uːj': 'ui'
+    'uj': 'ui'

data/maps/royin-tha-Thai-Latn-1999.yaml

@@ -0,0 +1,80 @@
+---
+authority_id: royin
+id: 1999
+language: iso-639-2:tha
+source_script: Thai
+destination_script: Latn
+name: Royal Thai General System of Transcription (1999)
+alias:
+  ogc11122:
+    code: tir_Thai2Latn_RIT_2000
+    description: Thai Royal Institute of Thailand 2000 System
+url: http://www.royin.go.th/wp-content/uploads/royin-ebook/276/FileUpload/758_6484.pdf
+creation_date: 1999
+adoption_date:
+description:
+
+notes: |
+
+tests:
+  - source: 'ภาษาไทย'
+    expected: 'phasathai'
+  - source: 'ไทย'
+    expected: 'thai'
+  - source: 'เชียงใหม่'
+    expected: 'chiangmai'
+  - source: 'ใหม่'
+    expected: 'mai'
+  - source: 'ใคร'
+    expected: 'khrai'
+  - source: "ที่"
+    expected: "thi"
+  - source: "เป็น"
+    expected: "pen"
+  - source: "ใน"
+    expected: "nai"
+  - source: "การ"
+    expected: "kan"
+  - source: "มี"
+    expected: "mi"
+  - source: "ได้"
+    expected: "dai"
+  - source: "ของ"
+    expected: "khong"
+  - source: "ไม่"
+    expected: "mai"
+  - source: "สถานีปางต้นผึ้ง"
+    expected: "sathanipangtonphueng"
+  - source: "ไพศาลี"
+    expected: "phaisali"
+  - source: "โรงเรียนไม้เรียงประชาสรรค์"
+    expected: "rongrianmairiangprachasan"
+
+map:
+  title-case: false
+  word_separator: " "
+  transcription: "sequitur.pythainlp_lexicon"
+
+  rules:
+
+  postrules:
+    - pattern: 'chh'
+      result: 'ch'
+
+  characters:
+    '[0-9]': ''
+
+  dictionary:
+    'ʰ': 'h'
+    'c': 'ch'
+    'ː': ''
+    'ŋ': 'ng'
+    'j': 'i'
+    'w': 'o'
+    'ɔ': 'o'
+    'ɤ': 'oe'
+    'ɛ': 'ae'
+    'ɯ': 'ue'
+    'ʔ' : ''
+    't͡ɕ': 'c'
+    '.' : ''

data/maps/{cn-chn-Hans-Latn-pinyin.yaml → sac-zho-Hans-Latn-1979.yaml}

@@ -1,18 +1,21 @@
 ---
-authority_id: cn
-id: pinyin
-language: chn
+authority_id: sac
+id: 1979
+language: iso-639-2:chn
 source_script: Hans
 destination_script: Latn
 name: Pinyin
+alias:
+  ogc11122:
+    code: zho_Hani2Latn_GCH_1979
+    description: Chinese Hanyu Pinyin 1979 System
 url: ftp://ftp.cuhk.hk/pub/chinese/ifcss/software/data/Uni2Pinyin.gz
 creation_date: 1996-04-17
 description: |
   This table contains the data on how Unicode Hanzi characters
   are pronounced in P.R.China.
 tests:
-  - source:
-    expected:
+
 
 notes:
   - Nine tab-separated columns. Column #1 is the Unicode (in hex), Columns #2 to #6 are Pinyin (tone '5' means Qingsheng)
@@ -13601,8 +13604,8 @@ map:
     "\u7A21": "zui4"
     "\u7A22": "yu4"
     "\u7A23": "su1"
-    "\u7A24": l��4
-    "\u7A25":
+    "\u7A24": "lue4"
+    "\u7A25": "xiang1"
     "\u7A26": "yi1"
     "\u7A27": "xi4"
     "\u7A28": "bian1"
@@ -21054,7 +21057,7 @@ map:
     "\u92DF": "qin2"
     "\u92E0": "shen4"
     "\u92E1": "han2"
-    "\u92E2": l��4
+    "\u92E2": "lue4"
     "\u92E3": "ye2"
     "\u92E4": "chu2"
     "\u92E5": "zeng4"