This is Python port of linkify-it.
Links recognition library with FULL unicode support. Focused on high quality link patterns detection in plain text.
Why it's awesome:
- Full unicode support, with astral characters!
- International domains support.
- Allows rules extension & custom normalizers.
pip install linkify-it-py
or
conda install -c conda-forge linkify-it-py
from linkify_it import LinkifyIt
linkify = LinkifyIt()
print(linkify.test("Site github.com!"))
# => True
print(linkify.match("Site github.com!"))
# => [linkify_it.main.Match({
# 'schema': '',
# 'index': 5,
# 'last_index': 15,
# 'raw': 'github.com',
# 'text': 'github.com',
# 'url': 'http://github.com'
# }]
from linkify_it import LinkifyIt
from linkify_it.tlds import TLDS
# Reload full tlds list & add unofficial `.onion` domain.
linkify = (
LinkifyIt()
.tlds(TLDS) # Reload with full tlds list
.tlds("onion", True) # Add unofficial `.onion` domain
.add("git:", "http:") # Add `git:` protocol as "alias"
.add("ftp:", None) # Disable `ftp:` protocol
.set({"fuzzy_ip": True}) # Enable IPs in fuzzy links (without schema)
)
print(linkify.test("Site tamanegi.onion!"))
# => True
print(linkify.match("Site tamanegi.onion!"))
# => [linkify_it.main.Match({
# 'schema': '',
# 'index': 5,
# 'last_index': 19,
# 'raw': 'tamanegi.onion',
# 'text': 'tamanegi.onion',
# 'url': 'http://tamanegi.onion'
# }]
from linkify_it import LinkifyIt
linkify = LinkifyIt()
def validate(obj, text, pos):
tail = text[pos:]
if not obj.re.get("twitter"):
obj.re["twitter"] = re.compile(
"^([a-zA-Z0-9_]){1,15}(?!_)(?=$|" + obj.re["src_ZPCc"] + ")"
)
if obj.re["twitter"].search(tail):
if pos > 2 and tail[pos - 2] == "@":
return False
return len(obj.re["twitter"].search(tail).group())
return 0
def normalize(obj, match):
match.url = "https://twitter.com/" + re.sub(r"^@", "", match.url)
linkify.add("@", {"validate": validate, "normalize": normalize})
Creates new linkifier instance with optional additional schemas.
By default understands:
http(s)://...
,ftp://...
,mailto:...
&//...
links- "fuzzy" links and emails (google.com, [email protected]).
schemas
is an dict, where each key/value describes protocol/rule:
- key - link prefix (usually, protocol name with
:
at the end,skype:
for example).linkify-it-py
makes sure that prefix is not preceded with alphanumeric char. - value - rule to check tail after link prefix
- str
- just alias to existing rule
- dict
- validate - either a
re.Pattern
(start with^
, and don't include the link prefix itself), or a validatorfunction
which, given arguments self, text and pos, returns the length of a match in text starting at index pos. pos is the index right after the link prefix. self can be used to access the linkify object to cache data. - normalize - optional function to normalize text & url of matched result (for example, for twitter mentions).
- validate - either a
- str
options
:
- fuzzy_link - recognize URL-s without
http(s)://
head. DefaultTrue
. - fuzzy_ip - allow IPs in fuzzy links above. Can conflict with some texts
like version numbers. Default
False
. - fuzzy_email - recognize emails without
mailto:
prefix. DefaultTrue
. - --- - set
True
to terminate link with---
(if it's considered as long dash).
Searches linkifiable pattern and returns True
on success or False
on fail.
Quick check if link MAY BE can exist. Can be used to optimize more expensive
.test()
calls. Return False
if link can not be found, True
- if .test()
call needed to know exactly.
Similar to .test()
but checks only specific protocol tail exactly at given
position. Returns length of found pattern (0 on fail).
Returns list
of found link matches or null if nothing found.
Each match has:
- schema - link schema, can be empty for fuzzy links, or
//
for protocol-neutral links. - index - offset of matched text
- last_index - index of next char after mathch end
- raw - matched text
- text - normalized text
- url - link, generated from matched text
Checks if a match exists at the start of the string. Returns Match
(see docs for match(text)
) or null if no URL is at the start.
Doesn't work with fuzzy links.
Load (or merge) new tlds list. Those are needed for fuzzy links (without schema) to avoid false positives. By default:
- 2-letter root zones are ok.
- biz|com|edu|gov|net|org|pro|web|xxx|aero|asia|coop|info|museum|name|shop|рф are ok.
- encoded (
xn--...
) root zones are ok.
If that's not enough, you can reload defaults with more detailed zones list.
Add a new schema to the schemas object. As described in the constructor
definition, key
is a link prefix (skype:
, for example), and value
is a str
to alias to another schema, or an dict
with validate
and
optionally normalize
definitions. To disable an existing rule, use
.add(key, None)
.
Override default options. Missed properties will not be changed.