ens-normalize 3.0.8

Ethereum Name Service (ENS) Name Normalizer

ENS Normalize Python

• Python implementation of ENSIP-15 - the ENS Name Normalization Standard.
  • Thanks to raffy.eth for his leadership in coordinating the definition of this standard with the ENS community.
  • This library is being maintained by the team at NameHash Labs as part of the greater NameGuard solution to help protect the ENS community.
• Passes 100% of the official validation tests (validated automatically with pytest on Linux, MacOS, and Windows, see below for details).
• Passes an additional suite of further tests for compatibility with the official Javascript reference implementation and code testing coverage.
• Based on JavaScript implementation version 1.9.0.

Glossary

Foundations

• sequence - a Unicode string containing any number of characters.
• label separator - a full stop character (also known as a period), e.g. . .
• label - a sequence of any length (including 0) that does not contain a label separator, e.g. abc or eth.
• name - a series of any number of labels (including 0) separated by label separators, e.g. abc.eth.

Names

• normalized name - a name that is in normalized form according to the ENS Normalization Standard. This means name == ens_normalize(name). A normalized name contains 0 or more labels. All labels in a normalized name always contain a sequence of at least 1 valid character. An empty string contains 0 labels and is a normalized name.
• normalizable name - a name that is normalized or that can be converted into a normalized name using ens_normalize.
• beautiful name - a name that is normalizable and is equal to itself when using ens_beautify. This means name == ens_beautify(name). For all normalizable names ens_normalize(ens_beautify(name)) == ens_normalize(name).
• disallowed name - a name that is not normalizable. This means ens_normalize(name) raises a DisallowedSequence.
• curable name - a name that is normalizable, or a name in the subset of disallowed names that can still be converted into a normalized name using ens_cure.
• empty name - a name that is the empty string. An empty string is a name with 0 labels. It is a normalized name.
• namehash ready name - a name that is ready for for use with the ENS namehash function. Only normalized names are namehash ready. Empty names represent the ENS namespace root for use with the ENS namehash function. Using the ENS namehash function on any name that is not namehash ready will return a node that is unreachable by ENS client applications that use a proper implementation of ens_normalize.

Sequences

• unnormalized sequence - a sequence from a name that is not in normalized form according to the ENS Normalization Standard.
• normalization suggestion - a sequence suggested as an in-place replacement for an unnormalized sequence.
• normalizable sequence - an unnormalized sequence containing a normalization suggestion that is automatically applied using ens_normalize and ens_cure.
• curable sequence - an unnormalized sequence containing a normalization suggestion that is automatically applied using ens_cure.
• disallowed sequence - an unnormalized sequence without any normalization suggestion.

The following Venn diagram is not to scale, but may help to communicate how some of the classifications of names relate to each other conceptually.

Usage

The package is available on pypi

pip install ens-normalize

You can also try it in Google Colab

ens_normalize

Normalize an ENS name:

from ens_normalize import ens_normalize
# str -> str
# raises DisallowedSequence for disallowed names
# output is namehash ready
ens_normalize('Nick.ETH')
# 'nick.eth'
# note: ens_normalize does not enforce any constraints that might be applied by a particular registrar. For example, the registrar for names that are a subname of '.eth' enforces a 3-character minimum and this constraint is not enforced by ens_normalize.

Check if a name is normalizable (see Glossary):

from ens_normalize import is_ens_normalizable
# str -> bool
is_ens_normalizable('Nick.ETH')
# True

Inspect issues with disallowed names:

from ens_normalize import DisallowedSequence, CurableSequence
try:
    # added a hidden "zero width joiner" character
    ens_normalize('Ni‍ck.ETH')
# Catch the first disallowed sequence (the name we are attempting to normalize could have more than one).
except DisallowedSequence as e:
    # error code
    print(e.code)
    # INVISIBLE

    # a message about why the sequence is disallowed
    print(e.general_info)
    # Contains a disallowed invisible character

    if isinstance(e, CurableSequence):
        # information about the curable sequence
        print(e.sequence_info)
        # 'This invisible character is disallowed'

        # starting index of the disallowed sequence in the input string
        # (counting in Unicode code points)
        print(e.index)
        # 2

        # the disallowed sequence
        # (use repr() to "see" the invisible character)
        print(repr(e.sequence))
        # '\u200d'

        # a normalization suggestion for fixing the disallowed sequence (there might be more disallowed sequences)
        print(repr(e.suggested))
        # ''
        # replacing the disallowed sequence with this suggestion (an empty string) represents the idea that the disallowed sequence is suggested to be removed

        # You may be able to fix this disallowed sequence by replacing e.sequence with e.suggested in the input string.
        # Fields index, sequence_info, sequence, and suggested are available only for curable errors.
        # Other disallowed sequences might be found even after applying this suggestion.

ens_cure

You can attempt conversion of disallowed names into normalized names using ens_cure. This algorithm can “cure” many normalization errors that would fail ens_normalize. This can be useful in some situations. For example, if a user input fails ens_normalize, a user could be prompted with a more helpful error message such as: “Did you mean curedname.eth?”.

Some names are not curable. For example, if it is challenging to provide a specific normalization suggestion that might be needed to replace a disallowed sequence.

Note: This function is NOT a part of the ENS Normalization Standard.

from ens_normalize import ens_cure
# input name with disallowed zero width joiner and '?'
# str -> str
ens_cure('Ni‍ck?.ETH')
# 'nick.eth'
# ZWJ and '?' are removed, no error is raised

# note: might remove all characters from the input, which would result in an empty name
ens_cure('?')
# '' (empty string)
# reason: '?' is disallowed and no replacement can be suggested

# note: might still raise DisallowedSequence for certain names, which cannot be cured, e.g.
ens_cure('0х0.eth')
# DisallowedSequence: Contains visually confusing characters from Cyrillic and Latin scripts
# reason: The "х" is actually a Cyrillic character that is visually confusing with the Latin "x".
#         However, the "0"s are standard Latin digits and it is not clear which characters should be removed.
#         They conflict with each other because it is not known if the user intended to use Cyrillic or Latin.

ens_beautify

Get a beautiful name that is optimized for display:

from ens_normalize import ens_beautify
# works like ens_normalize()
# output ready for display
ens_beautify('1⃣2⃣.eth')
# '1️⃣2️⃣.eth'

# note: normalization is unchanged:
# ens_normalize(ens_beautify(x)) == ens_normalize(x)
# note: in addition to beautifying emojis with fully-qualified emoji, ens_beautify converts the character 'ξ' (Greek lowercase 'Xi') to 'Ξ' (Greek uppercase 'Xi', a.k.a. the Ethereum symbol) in labels that contain no other Greek characters

ens_tokenize

Generate detailed name analysis:

from ens_normalize import ens_tokenize
# str -> List[Token]
# always returns a tokenization of the input
ens_tokenize('Nàme‍🧙‍♂.eth')
# [TokenMapped(cp=78, cps=[110], type='mapped'),
#  TokenNFC(input=[97, 768], cps=[224], type='nfc'),
#  TokenValid(cps=[109, 101], type='valid'),
#  TokenDisallowed(cp=8205, type='disallowed'),
#  TokenEmoji(emoji=[129497, 8205, 9794, 65039],
#             input=[129497, 8205, 9794],
#             cps=[129497, 8205, 9794],
#             type='emoji'),
#  TokenStop(cp=46, type='stop'),
#  TokenValid(cps=[101, 116, 104], type='valid')]

ens_normalizations

For a normalizable name, you can find out how the input is transformed during normalization:

from ens_normalize import ens_normalizations
# Returns a list of transformations (unnormalized sequence -> normalization suggestion)
# that have been applied to the input during normalization.
# NormalizableSequence has the same fields as CurableSequence:
# - code
# - general_info
# - sequence_info
# - index
# - sequence
# - suggested
ens_normalizations('Nàme🧙‍♂️.eth')
# [NormalizableSequence(code="MAPPED", index=0, sequence="N", suggested="n"),
#  NormalizableSequence(code="FE0F", index=4, sequence="🧙‍♂️", suggested="🧙‍♂")]

Example Workflow

An example normalization workflow:

name = 'Nàme🧙‍♂️.eth'
try:
    normalized = ens_normalize(name)
    print('Normalized:', normalized)
    # Normalized: nàme🧙‍♂.eth
    # Success!

     # was the input transformed by the normalization process?
    if name != normalized:
        # Let's check how the input was changed:
        for t in ens_normalizations(name):
            print(repr(t)) # use repr() to print more information
        # NormalizableSequence(code="MAPPED", index=0, sequence="N", suggested="n")
        # NormalizableSequence(code="FE0F", index=4, sequence="🧙‍♂️", suggested="🧙‍♂")
        #                                     invisible character inside emoji ^
except DisallowedSequence as e:
    # Even if the name is invalid according to the ENS Normalization Standard,
    # we can try to automatically cure disallowed sequences.
    try:
        print('Cured:', ens_cure(name))
    except DisallowedSequence as e:
        # The name cannot be automatically cured.
        print('Disallowed name error:', e)

You can run many of the above functions at once. It is faster than running all of them sequentially.

from ens_normalize import ens_process
# use only the do_* flags you need
ens_process("Nàme🧙‍♂️1⃣.eth",
    do_normalize=True,
    do_beautify=True,
    do_tokenize=True,
    do_normalizations=True,
    do_cure=True,
)
# ENSProcessResult(
#   normalized='nàme🧙\u200d♂1⃣.eth',
#   beautified='nàme🧙\u200d♂️1️⃣.eth',
#   tokens=[...],
#   cured='nàme🧙\u200d♂1⃣.eth',
#   cures=[], # This is the list of cures that were applied to the input (in this case, none).
#   error=None, # This is the exception raised by ens_normalize().
#               # It is a DisallowedSequence or CurableSequence if the error is curable.
#   normalizations=[
#     NormalizableSequence(code="MAPPED", index=0, sequence="N", suggested="n"),
#     NormalizableSequence(code="FE0F", index=4, sequence="🧙‍♂️", suggested="🧙‍♂")
#   ])

Exceptions

These Python classes are used by the library to communicate information about unnormalized sequences.

Exception class	ens_normalize handling	ens_cure handling	normalization suggestion	Inherits From
NormalizableSequence	✅ automatically resolves	✅ automatically resolves	✅ included	CurableSequence
CurableSequence	❌ throws error	✅ automatically resolves	✅ included	DisallowedSequence
DisallowedSequence	❌ throws error	❌ throws error	❌ none	Exception

List of all NormalizableSequence types

NormalizableSequenceType	General info	Sequence info
IGNORED	Contains a disallowed "ignored" character that has been removed	This character is ignored during normalization and has been automatically removed
MAPPED	Contains a disallowed character that has been replaced by a normalized sequence	This character is disallowed and has been automatically replaced by a normalized sequence
FE0F	Contains a disallowed variant of an emoji which has been replaced by an equivalent normalized emoji	This emoji has been automatically fixed to remove an invisible character
NFC	Contains a disallowed sequence that is not "NFC normalized" which has been replaced by an equivalent normalized sequence	This sequence has been automatically normalized into NFC canonical form

List of all CurableSequence types

Curable errors contain additional information about the disallowed sequence and a normalization suggestion that might help to cure the name.

CurableSequenceType	General info	Sequence info
UNDERSCORE	Contains an underscore in a disallowed position	An underscore is only allowed at the start of a label
HYPHEN	Contains the sequence '--' in a disallowed position	Hyphens are disallowed at the 2nd and 3rd positions of a label
EMPTY_LABEL	Contains a disallowed empty label	Empty labels are not allowed, e.g. abc..eth
CM_START	Contains a combining mark in a disallowed position at the start of the label	A combining mark is disallowed at the start of a label
CM_EMOJI	Contains a combining mark in a disallowed position after an emoji	A combining mark is disallowed after an emoji
DISALLOWED	Contains a disallowed character	This character is disallowed
INVISIBLE	Contains a disallowed invisible character	This invisible character is disallowed
FENCED_LEADING	Contains a disallowed character at the start of a label	This character is disallowed at the start of a label
FENCED_MULTI	Contains a disallowed consecutive sequence of characters	Characters in this sequence cannot be placed next to each other
FENCED_TRAILING	Contains a disallowed character at the end of a label	This character is disallowed at the end of a label
CONF_MIXED	Contains visually confusing characters from multiple scripts ({script1}/{script2})	This character from the {script1} script is disallowed because it is visually confusing with another character from the {script2} script

List of all DisallowedSequence types

Disallowed name errors are not considered curable because it may be challenging to suggest a specific normalization suggestion that might resolve the problem.

DisallowedSequenceType	General info	Explanation
NSM_REPEATED	Contains a repeated non-spacing mark	Non-spacing marks can be encoded as one codepoint with the preceding character, which makes it difficult to suggest a normalization suggestion
NSM_TOO_MANY	Contains too many consecutive non-spacing marks	Non-spacing marks can be encoded as one codepoint with the preceding character, which makes it difficult to suggest a normalization suggestion
CONF_WHOLE	Contains visually confusing characters from {script1} and {script2} scripts	Both characters are equally likely to be the correct character to use and a normalization suggestion cannot be provided

Development

Update this library to the latest ENS normalization specification (optional)

This library uses files defining the normalization standard directly from the official Javascript implementation. When the standard is updated with new characters, this library can be updated by running the following steps:

1. Requirements:

   • Node.js >= 18
   • npm

2. Set the hash of the latest commit from the JavaScript library inside package.json

3. Run the updater:

   cd tools/updater
   npm start
   

Build and test

Installs dependencies, runs validation tests and builds the wheel.

1. Install requirements:

   • Python
   • Poetry

2. Install dependencies:

   poetry install
   

3. Run tests (including official validation tests):

   poetry run pytest
   

4. Build Python wheel:

   poetry build
   

License

Licensed under the MIT License, Copyright © 2023-present NameHash Labs.

See LICENSE for more information.