data version 0.3-1.0.4

analysis/ummama.ipynb

@@ -516,7 +516,8 @@
     "introNouns = collections.Counter()\n",
     "\n",
     "for (line, um, ma1, word, ma2) in results:\n",
-    "  introNouns[F.symr.v(word)] += 1\n",
+    "  strippedWord = L.d(word, otype='sign')[:-1]\n",
+    "  introNouns[F.symr.v(strippedWord)] += 1\n",
     "\n",
     "len(introNouns)"
    ]

characters/mapping.tsv

@@ -8,6 +8,7 @@
 1(esze3)	𒑘
 1(gesz'u)	𒐞
 1(gesz2)	𒐕
+1(iku)	𒀸
 1(szar2)	𒊹
 1(u)	𒌋
 1/2(disz)	𒈦
@@ -23,6 +24,7 @@
 2(gesz'u)	𒐟
 2(gesz2)	𒐖
 2(gisz)	𒄑
+2(iku)	𒐀
 2(szar2)	𒐣
 2(u)	𒌋
 2/3(disz)	𒑛
@@ -32,8 +34,10 @@
 3(bur'u)	𒐶
 3(bur3)	𒌋𒌋𒌋
 3(disz)	𒐈
+3(esze3)	𒀸𒌋
 3(gesz'u)	𒐠
 3(gesz2)	𒐗
+3(iku)	𒐁
 3(u)	𒌋
 4(asz)	𒐂
 4(ban2)	𒑒
@@ -43,6 +47,7 @@
 4(disz)	𒐉
 4(gesz'u)	𒐡
 4(gesz2)	𒐘
+4(iku)	𒐂
 4(u)	𒐏
 5(asz)	𒐃
 5(ban2)	𒑔
@@ -51,24 +56,29 @@
 5(bur3)	𒐐
 5(disz)	𒐊
 5(gesz2)	𒐙
+5(iku)	𒐃
 5(u)	𒐐
 5/6(disz)	𒑜
 6(asz)	𒐄
 6(bur3)	𒐑
 6(disz)	𒐋
 6(gesz2)	𒐚
+6(iku)	𒐄
 7(asz)	𒐅
 7(bur3)	𒐒
 7(disz)	𒐌
 7(gesz2)	𒐛
+7(iku)	𒐅
 8(asz)	𒐆
 8(bur3)	𒐓
 8(disz)	𒐍
 8(gesz2)	𒐜
+8(iku)	𒐆
 9(asz)	𒐇
 9(bur3)	𒐔
 9(disz)	𒐎
 9(gesz2)	𒐝
+9(iku)	𒐇
 A	𒀀
 AB	𒀊
 AD	𒀜
@@ -366,6 +376,7 @@ el3	𒀭
 elam	𒉏
 em	𒅎
 eme	𒅴
+eme6	𒀲𒊩
 en	𒂗
 en6	𒅔
 engar	𒀳

docs/transcription.md

@@ -98,7 +98,7 @@ There are several types of sign, stored in the feature `type`.
 type | examples | description
 ------- | ------ | ------
 `reading` | `ma` `qa2` | normal sign with a reading (lowercase)
-`unknown` | `x` | representation of an unknown sign
+`unknown` | `x` `n` | representation of an unknown sign, the `n` stands for an unknown numeral
 `numeral` | `5(disz)` `5/6(disz)`  | a numeral, either with a repeat or with a fraction
 `ellipsis` | `...` | representation of an unknown number of missing signs
 `grapheme` | `ARAD2` `GAN2` | sign given as a grapheme (uppercase)
@@ -291,11 +291,16 @@ Simple signs may be *augmented* with *flags* (see below).
 
 ### Unknown signs ###
 
-The letters `x` and `X` in isolation stand for an unknown signs.
+The letters `x` and `X`, `n` and `N` in isolation stand for an unknown signs.
 
 The *type* of such signs is `unknown`. 
 
-If the value is `x`, it will stored in **reading**, if it is `X` in **grapheme**. 
+If the value is `x` or `n`, it will stored in **reading**, if it is `X` or `N` in **grapheme**.
+
+The `x` and `X` stand for completely unknown signs, the `n` and `N` stand for unknown signs
+of which it is known that they are numerals.
+
+**N.B:** See under numerals below, where `n` plays a slightly different role.
 
 ### Ellipsis ###
 
@@ -318,7 +323,8 @@ Numeric signs may also be preceded with a *fraction*:
 We store the integral number before the brackets in the feature **repeat**,
 and the fraction in the feature **fraction**.
 
-If the repeat is `n`, it means that a number is missing.
+If the repeat is `n`, it means that the amount of repetition is uncertain
+or that a repetition is missing.
 We store it as `repeat` = `-1`, so repeats always have an integer value.
 
 In a numeral, within the brackets you find the **reading** or **grapheme**,

programs/tfFromATF.py

@@ -13,8 +13,8 @@
 BASE = os.path.expanduser('~/github')
 ORG = 'Nino-cunei'
 REPO = 'oldbabylonian'
-VERSION_SRC = '0.2'
-VERSION_TF = '1.0.3'
+VERSION_SRC = '0.3'
+VERSION_TF = '1.0.4'
 REPO_DIR = f'{BASE}/{ORG}/{REPO}'
 
 TRANS_DIR = f'{REPO_DIR}/sources/cdli/transcriptions'
@@ -28,17 +28,9 @@
 TF_DIR = f'{REPO_DIR}/tf'
 OUT_DIR = f'{TF_DIR}/{VERSION_TF}'
 
-# SOURCE FIXES
-
-SRC_FIXES = dict(
-    bub='dub',
-    umi='um i',
-    ura='ra',
-    szii='szi',
-)
 #  CHARACTERS
 
-UNMAPPABLE = {'x', 'X', '...'}
+UNMAPPABLE = {'x', 'X', 'n', 'N', '...'}
 
 prime = "'"
 ellips = '…'
@@ -51,7 +43,7 @@
     't,': 'ţ',
 }
 
-unknownStr = 'xX'
+unknownStr = 'xXnN'
 unknownSet = set(unknownStr)
 
 lowerLetterStr = 'abcdefghijklmnopqrstuvwyz' + ''.join(emphatic.values())
@@ -657,12 +649,6 @@ def getMapping():
 
 
 def getSources():
-  if os.path.exists(OUT_DIR):
-    rmtree(OUT_DIR)
-  os.makedirs(OUT_DIR, exist_ok=True)
-
-# list all sources
-
   return tuple(
       os.path.splitext(os.path.basename(f))[0]
       for f in glob(f'{IN_DIR}/*.txt')
@@ -693,6 +679,11 @@ def checkSane(line):
 
 
 def convert():
+  if generateTf:
+    if os.path.exists(OUT_DIR):
+      rmtree(OUT_DIR)
+    os.makedirs(OUT_DIR, exist_ok=True)
+
   cv = getConverter()
 
   return cv.walk(
@@ -1708,11 +1699,6 @@ def getParts(word):
         if isNumbered:
           ln = isNumbered.group(1)
           recentTrans = isNumbered.group(2)
-          for (pat, rep) in SRC_FIXES.items():
-            newRecentTrans = recentTrans.replace(pat, rep)
-            if newRecentTrans != recentTrans:
-              warnings[f'source fix: {pat} => {rep}'][src].add((i, line, pNum, None))
-              recentTrans = newRecentTrans
 
         else:
           errors[f'line: not numbered'][src].add((i, line, pNum, None))