v0.1.4

All dependencies relative to hash function implementation are now optional, dynamically imported only when the hash functions using them are used for the first time. Running `pip install --upgrade multiformats` will not install any of them, but they can be all installed by running `pip install --upgrade multiformats[full]`. In particular, this closes #4.

Hash function implementations are loaded and registered transparently on first use, to reduce memory footprint and module loading times. Analogously, a number of multibases are created and registered transparently on first use.

All hash functions with a readily available, well-supported existing Python implementation are now supported.

Finally, closes #3.

Author: sg495

MULTIFORMATS-LICENSE renamed to ADDITIONAL-LICENSES

@@ -1,8 +1,10 @@
-The following items are subject to MIT License by Protocol Labs Inc:
+The following items are subject to MIT License by Protocol Labs Inc, included below:
 
 - multibase table, downloaded from https://github.com/multiformats/multibase/raw/master/multibase.csv
 - multicodec table, downloaded from https://github.com/multiformats/multicodec/raw/master/table.csv
-- test vectors for multihash, downloaded from https://github.com/multiformats/multihash/raw/master/tests/values/test_cases.csv on 14 Dec 2021
+- the test vectors for multihash in multihash-test-str-vectors.csv, downloaded from https://github.com/multiformats/multihash/raw/master/tests/values/test_cases.csv on 14 Dec 2021
+
+Test vectors for murmur3 hash are public domain, courtesy of Ian Boyd https://stackoverflow.com/questions/14747343/murmurhash3-test-vectors#31929528
 
 
 The MIT License (MIT)

README.rst

@@ -44,6 +44,27 @@ You can install the latest release from `PyPI <https://pypi.org/project/multifor
 
     $ pip install --upgrade multiformats
 
+The following are mandatory dependencies for this module:
+
+- `typing-extensions <https://github.com/python/typing_extensions>`_, for backward compatibility of static typing.
+- `typing-validation <https://github.com/hashberg-io/typing-validation>`_, for dynamic typechecking
+- `bases <https://github.com/hashberg-io/bases>`_, for implementation of base encodings used by Multibase
+
+The following are optional dependencies for this module:
+
+- `pysha3 <https://github.com/tiran/pysha3>`_, for the ``keccak`` hash functions.
+- `blake3 <https://github.com/oconnor663/blake3-py>`_, for the ``blake3`` hash function.
+- `pyskein <https://pythonhosted.org/pyskein/>`_, for the ``skein`` hash functions.
+- `mmh3 <https://github.com/hajimes/mmh3>`_, for the ``murmur3`` hash functions.
+- `pycryptodomex <https://github.com/Legrandin/pycryptodome/>`_, for the ``ripemd-160`` hash function, \
+  the ``kangarootwelve`` hash function and the ``sha2-512-224``/``sha2-512-256`` hash functions.
+
+You can install the latest release together with all optional dependencies as follows:
+
+.. code-block:: console
+
+    $ pip install --upgrade multiformats[full]
+
 
 Usage
 -----
@@ -311,3 +332,5 @@ License
 -------
 
 `MIT © Hashberg Ltd. <LICENSE>`_
+
+See `additional Licenses <ADDITIONAL-LICENSES>`_ for licensing of the multicodec table, the multibase table and test vectors for multihashes.

docs/api/multiformats.multihash.raw.rst

@@ -6,7 +6,8 @@ multiformats.multihash.raw
 Hashfun
 -------
 
-.. autodata:: multiformats.multihash.raw.Hashfun
+.. autoclass:: multiformats.multihash.raw.Hashfun
+    :members:
 
 MultihashImpl
 -------------

docs/getting-started.rst

@@ -29,4 +29,25 @@ The above will import the following names:
 The first five are modules implementing the homonymous specifications,
 while :class:`~multiformats.cid.CID` is a class for Content IDentifiers.
 
+The following are mandatory dependencies for this module:
+
+- `typing-extensions <https://github.com/python/typing_extensions>`_, for backward compatibility of static typing.
+- `typing-validation <https://github.com/hashberg-io/typing-validation>`_, for dynamic typechecking
+- `bases <https://github.com/hashberg-io/bases>`_, for implementation of base encodings used by Multibase
+
+The following are optional dependencies for this module:
+
+- `pysha3 <https://github.com/tiran/pysha3>`_, for the ``keccak`` hash functions.
+- `blake3 <https://github.com/oconnor663/blake3-py>`_, for the ``blake3`` hash function.
+- `pyskein <https://pythonhosted.org/pyskein/>`_, for the ``skein`` hash functions.
+- `mmh3 <https://github.com/hajimes/mmh3>`_, for the ``murmur3`` hash functions.
+- `pycryptodomex <https://github.com/Legrandin/pycryptodome/>`_, for the ``ripemd-160`` hash function, \
+  the ``kangarootwelve`` hash function and the ``sha2-512-224``/``sha2-512-256`` hash functions.
+
+You can install the latest release together with all optional dependencies as follows:
+
+.. code-block:: console
+
+    $ pip install --upgrade multiformats[full]
+
 GitHub repo: https://github.com/hashberg-io/multiformats

docs/make-api.json

@@ -11,10 +11,12 @@
     "exclude_members": {
         "multiformats.multicodec": ["build_multicodec_tables"],
         "multiformats.multibase": ["build_multibase_tables"],
-        "multiformats.multibase.raw": ["identity_raw_encoder", "identity_raw_decoder", "proquint_raw_encoder", "proquint_raw_decoder", "RawEncoder", "RawDecoder"],
+        "multiformats.multibase.raw": ["RawEncoder", "RawDecoder"],
         "multiformats.cid": ["CIDVersionNumbers", "byteslike"],
         "multiformats.multiaddr.raw": ["ip4_encoder", "ip4_decoder", "ip6_encoder", "ip6_decoder", "tcp_udp_encoder", "tcp_udp_decoder"]
     },
     "include_modules": [],
-    "exclude_modules": []
-}
+    "exclude_modules": [
+        "multiformats.multihash._hashfuns"
+    ]
+}

docs/make-api.py

@@ -87,8 +87,12 @@ def make_apidocs() -> None:
         os.remove(apidoc_file)
     print()
 
+    mod_name_to_del: List[str] = []
+
     for mod_name, mod in modules_dict.items():
-        if mod_name in exclude_modules:
+        if any(mod_name.startswith(name) for name in exclude_modules):
+        # if mod_name in exclude_modules:
+            mod_name_to_del.append(mod_name)
             continue
         filename = f"{apidocs_folder}/{mod_name}.rst"
         print(f"Writing API docfile {filename}")
@@ -164,6 +168,10 @@ def make_apidocs() -> None:
             f.write("\n".join(lines))
         print("")
 
+
+    for mod_name in mod_name_to_del:
+        del modules_dict[mod_name]
+
     toctable_lines = [
         ".. toctree::",
         "    :maxdepth: 2",

docs/requirements.txt

@@ -4,4 +4,3 @@ sphinx_autodoc_typehints
 bases
 typing-extensions
 typing-validation
-pyskein

multiformats/__init__.py

@@ -15,7 +15,7 @@
     while :class:`~multiformats.cid.CID` is a class for Content IDentifiers.
 """
 
-__version__ = "0.1.3"
+__version__ = "0.1.4"
 
 from . import varint
 from . import multicodec

multiformats/multiaddr/err.py

@@ -6,8 +6,7 @@
 
 class MultiaddrKeyError(builtins.KeyError): # pylint: disable = redefined-builtin
     """ Class for :mod:`~multiformats.multiaddr` key errors. """
-    ...
+
 
 class MultiaddrValueError(builtins.ValueError): # pylint: disable = redefined-builtin
     """ Class for :mod:`~multiformats.multiaddr` value errors. """
-    ...

multiformats/multibase/__init__.py

@@ -39,7 +39,7 @@ class Multibase:
 
         :param name: the multibase name
         :type name: :obj:`str`
-        :param code: the multibase code, as single-char string or ``0xYZ`` hex-string of a byte
+        :param code: the multibase code, as single-char string or ``0x...`` hex-string of a non-empty bytestring
         :type code: :obj:`str`
         :param status: the multibase status
         :type status: ``'draft'``, ``'candidate'`` or ``'default'``, *optional*
@@ -91,20 +91,20 @@ def validate_code(code: str) -> str:
             MultibaseValueError: Multibase codes must be single-character strings
             or the hex digits '0xYZ' of a single byte.
 
-            :param code: the multibase code, as single character or ``0xYZ`` hex-string of a single byte
+            :param code: the multibase code, as single character or ``0x...`` hex-string of a non-empty bytestring
             :type code: :obj:`str`
 
             :raises ValueError: if the code is invalid
 
         """
         validate(code, str)
-        if re.match(r"^0x[0-9a-zA-Z][0-9a-zA-Z]$", code):
+        if re.match(r"^0x([0-9a-zA-Z][0-9a-zA-Z])+$", code):
             ord_code = int(code, base=16)
+            if ord_code in range(0x20, 0x7F):
+                raise MultibaseValueError("Multibase codes in hex format cannot be printable ASCII characters.")
             code = chr(ord_code)
         elif len(code) != 1:
-            raise MultibaseValueError("Multibase codes must be single-character strings or the hex digits '0xYZ' of a single byte.")
-        if ord(code) not in range(0x00, 0x80):
-            raise MultibaseValueError("Multibase codes must be ASCII characters.")
+            raise MultibaseValueError("Multibase codes must be single-character strings or the hex digits '0x...' of a non-empty bytestring.")
         return code
 
     @staticmethod
@@ -145,7 +145,9 @@ def code_printable(self) -> str:
         code = self.code
         ord_code = ord(code)
         if ord_code not in range(0x20, 0x7F):
-            return "0x"+base16.encode(bytes([ord_code]))
+            ord_code_num_bytes = max(1, math.ceil(ord_code.bit_length()/8))
+            ord_code_bytes = ord_code.to_bytes(ord_code_num_bytes, byteorder="big")
+            return "0x"+base16.encode(ord_code_bytes)
         return code
 
     @property
@@ -555,6 +557,6 @@ def build_multibase_tables(bases: Iterable[Multibase]) -> Tuple[Dict[str, Multib
 # Create the global code->multibase and name->multibase mappings.
 _code_table: Dict[str, Multibase]
 _name_table: Dict[str, Multibase]
-with importlib_resources.open_text("multiformats.multibase", "multibase-table.json") as _table_f:
+with importlib_resources.open_text("multiformats.multibase", "multibase-table.json", encoding="utf8") as _table_f:
     _table_json = json.load(_table_f)
     _code_table, _name_table = build_multibase_tables(Multibase(**row) for row in _table_json)