Fixed codec: base122

dhondta · dhondta · commit e443fd408118 · 2021-10-02T11:39:55.000+02:00
diff --git a/codext/base/base122.py b/codext/base/base122.py
@@ -10,16 +10,88 @@
 from ..__common__ import *
 
 
-# no __examples__ ; handled manually in tests/test_base.py
+__examples__ = {
+    'enc(base122|base-122)': {
+        'this is a test': ":\x1aʗ\x19\x01Rs\x10\x18$\x07#\x15ft",
+        b'This is another longer test string with d1g1t5 and sp3c141 characters !\n': \
+            b"*\x1a\xca\x97\x19\x01Rs\x10\x18-f{QPe9\x08\xcb\x86{9Ne9\x08\x0eF+Mh 9]\x0e\xd3\x8b"
+            b"9N ;Z.FA\x01H13L.C)\x01Bn2\x08\x0e7\x01MF1\x1a\x0c$\x06\x1b!Br0XnF+If \x10B@"
+    },
+    'enc-dec(base_122)': ["@random"],
+}
 
 
-def base122_encode(input, errors="strict"):
-    raise NotImplementedError
+_BAD = [0, 10, 13, 34, 38, 92]
+_i = lambda c: c if isinstance(c, int) else ord(c)
 
 
-def base122_decode(input, errors="strict"):
-    raise NotImplementedError
+if PY3:
+    # inspired from: https://github.com/kevinAlbs/Base122/blob/master/base122.js
+    def base122_encode(input, errors="strict"):
+        idx, bit, r, l = 0, 0, [], len(input)
+        
+        def _get_7bits(idx, bit):
+            if idx >= l:
+                return idx, bit, False
+            B1 = _i(input[idx])
+            p1 = (((254 >> bit) & B1) << bit) >> 1
+            bit += 7
+            if bit < 8:
+                return idx, bit, p1
+            bit -= 8
+            idx += 1
+            if idx >= l:
+                return idx, bit, p1
+            B2 = _i(input[idx])
+            p2 = (((65280 >> bit) & B2) & 255) >> (8 - bit)
+            return idx, bit, (p1 | p2)
+        
+        while True:
+            if idx >= l:
+                break
+            # get seven bits of input data
+            idx, bit, B = _get_7bits(idx, bit)
+            # check for illegal chars
+            try:
+                bad_idx = _BAD.index(B)
+            except ValueError:
+                r.append(B)
+                continue
+            idx, bit, nB = _get_7bits(idx, bit)
+            if nB is False:
+                nB, bad_idx = B, 7
+            B1, B2 = 194, 128
+            B1 |= (7 & bad_idx) << 2
+            B1 |= int((nB & 64) > 0)
+            B2 |= nB & 63
+            r.extend([B1, B2])
+        return "".join(map(chr, r)).encode("latin-1"), len(input)
 
 
-add("base122", base122_encode, base122_decode, r"^base[-_]?122$")
+    # inspired from: https://github.com/kevinAlbs/Base122/blob/master/base122.js
+    def base122_decode(input, errors="strict"):
+        currB, bob, r, input = 0, 0, [], list(map(ord, input))
+        
+        def _get_7bits(currB, bob, B, decoded):
+            B <<= 1
+            currB |= (B % 0x100000000) >> bob
+            bob += 7
+            if bob >= 8:
+                decoded += [currB]
+                bob -= 8
+            return (B << (7 - bob)) & 255, bob
+
+        for i in range(len(input)):
+            if input[i] >= 128:
+                try:
+                    currB, bob = _get_7bits(currB, bob, _BAD[(input[i] >> 8) & 7], r)
+                except IndexError:
+                    pass
+                currB, bob = _get_7bits(currB, bob, input[i] & 127, r)
+            else:
+                currB, bob = _get_7bits(currB, bob, input[i], r)
+        return "".join(map(chr, r)), len(input)
+
+
+    add("base122", base122_encode, base122_decode, r"^base[-_]?122$")
 
diff --git a/docs/enc/base.md b/docs/enc/base.md
@@ -134,8 +134,7 @@ This encoding relies on the `base64` library and is only supported in Python 3.
 :---: | :---: | --- | ---
 `base85` | text <-> Base85 encoded text | `base[-_]?85` | Python 3 only (relies on `base64` module)
 `base100` | text <-> Base100 encoded text | `base[-_]?100|emoji` | Python 3 only
-
-With Python 3, `base85` and `base100` (emoji's) are also supported.
+`base122` | text <-> Base122 encoded text | `base[-_]?122` | Python 3 only
 
 ```python
 >>> codecs.encode("this is a test", "base85")
@@ -151,3 +150,10 @@ With Python 3, `base85` and `base100` (emoji's) are also supported.
 'this is a test'
 ```
 
+```python
+>>> codecs.encode("this is a test", "base122")
+':\x1aʗ\x19\x01Rs\x10\x18$\x07#\x15ft'
+>>> codecs.decode(":\x1aʗ\x19\x01Rs\x10\x18$\x07#\x15ft", "base122")
+'this is a test'
+```
+
diff --git a/docs/howto.md b/docs/howto.md
@@ -192,7 +192,7 @@ __examples__ = {
 
 ```python
 __examples__ = {
-    'enc-dec(codec)': ["test string", "TEST STRING"]
+    'enc-dec(codec)': ["test string", "TEST STRING", "@random", "@random{1024}"]
 }
 ```
 
@@ -201,6 +201,7 @@ __examples__ = {
     - `__examples__` is the standard dunder, thus not specified in `add`/`add_map`.
     - `enc-dec` is used, meaning that a list of inputs is defined.
     - So, whatever its encoded output, the input string shall give the same while applying encoding then decoding.
+    - The special values `@random` and `@random{1024}`, meaning that test strings are generated from any possible byte-character with a specified length (512 when not specified, otherwise specified with `{...}`).
 
 ```python
 __examples__ = {
