302 lines
18 KiB
Plaintext
302 lines
18 KiB
Plaintext
|
Ë
|
|||
|
|
vR j5C ã óÎ — d dl d dlmZmZmZ ddlmZ ddlZej dk\ Z G d„ d« Z e
|
|||
|
|
e « j « D � cg c] } e | e
|
|||
|
|
« sŒ
|
|||
|
|
| ‘Œ c} Zyc c} w ) é )Ú*)Ú
|
|||
|
|
DelimitedListÚ any_open_tagÚ
|
|||
|
|
any_close_tagé )ÚdatetimeN)é é
|
|||
|
|
c ó
|
|||
|
|
— e Zd ZdZed„ « Zed„ « Z ee« j d« j e rend„ « Z ee
|
|||
|
|
« j d« j eed« « Z ed« j d « j e rend
|
|||
|
|
„ « Z e« j e rend „ « d z e« j e rend
|
|||
|
|
„ « z j d« Z ej) d„ « ee e ed« j- « ez « z z j d« Z ej) e« ed« j d« j e rend„ « Z ed« j d« j e rend„ « Z eez ez j d« j7 « Z ed« j d« j e rend„ « Z ed « j d « j e rend „ « Z ee e « j d « Z! ed « j d!« Z" ed"« j d#« Z#e#d$e#z d%z z j d&« Z$ ee#d$e#z d'z z « d(z ee#d$e#z d'z z « z j d)« Z%e%jM d*„ « d+e"z j d,« Z' e(e$e'z e%z j d-« « j d-« Z) ed.« j d/« Z* edQd0e+fd1„« Z,edRd0e+fd2„« Z- ed3« j d4« Z. ed5« j d6« Z/ ed7„ « Z0e r. e.« j) e0« Z1 e/« j) e0« Z2 ed8« j d9« Z3 e4j, « e5j, « z Z6ed:e+d;ed<e7fd=„« Z8 e( e9 e:d>« e;« z ee<d>¬?« z e e=d@« e> e;« d>z « z « z « « j7 « j dA« Z? e@ eeAj… « e?z dB¬C« « j dD« ZC edE„ « ZDedF„ « ZE edG« j dH« ZF e eGdIe« « ZH e eGdJe« « ZI e eGdKe,« « ZJ e eGdLe-« « ZK e eGdMe8« « ZL e eGdNeD« « ZM e eGdOeE« « ZNyP)SÚpyparsing_commona
|
|||
|
|
Here are some common low-level expressions that may be useful in
|
|||
|
|
jump-starting parser development:
|
|||
|
|
|
|||
|
|
- numeric forms (:class:`integers<integer>`, :class:`reals<real>`,
|
|||
|
|
:class:`scientific notation<sci_real>`)
|
|||
|
|
- common :class:`programming identifiers<identifier>`
|
|||
|
|
- network addresses (:class:`MAC<mac_address>`,
|
|||
|
|
:class:`IPv4<ipv4_address>`, :class:`IPv6<ipv6_address>`)
|
|||
|
|
- ISO8601 :class:`dates<iso8601_date>` and
|
|||
|
|
:class:`datetime<iso8601_datetime>`
|
|||
|
|
- :class:`UUID<uuid>`
|
|||
|
|
- :class:`comma-separated list<comma_separated_list>`
|
|||
|
|
- :class:`url`
|
|||
|
|
|
|||
|
|
Parse actions:
|
|||
|
|
|
|||
|
|
- :class:`convert_to_integer`
|
|||
|
|
- :class:`convert_to_float`
|
|||
|
|
- :class:`convert_to_date`
|
|||
|
|
- :class:`convert_to_datetime`
|
|||
|
|
- :class:`strip_html_tags`
|
|||
|
|
- :class:`upcase_tokens`
|
|||
|
|
- :class:`downcase_tokens`
|
|||
|
|
|
|||
|
|
Examples:
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
pyparsing_common.number.run_tests('''
|
|||
|
|
# any int or real number, returned as the appropriate type
|
|||
|
|
100
|
|||
|
|
-100
|
|||
|
|
+100
|
|||
|
|
3.14159
|
|||
|
|
6.02e23
|
|||
|
|
1e-12
|
|||
|
|
''')
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
:options: +NORMALIZE_WHITESPACE
|
|||
|
|
|
|||
|
|
|
|||
|
|
# any int or real number, returned as the appropriate type
|
|||
|
|
100
|
|||
|
|
[100]
|
|||
|
|
|
|||
|
|
-100
|
|||
|
|
[-100]
|
|||
|
|
|
|||
|
|
+100
|
|||
|
|
[100]
|
|||
|
|
|
|||
|
|
3.14159
|
|||
|
|
[3.14159]
|
|||
|
|
|
|||
|
|
6.02e23
|
|||
|
|
[6.02e+23]
|
|||
|
|
|
|||
|
|
1e-12
|
|||
|
|
[1e-12]
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
pyparsing_common.fnumber.run_tests('''
|
|||
|
|
# any int or real number, returned as float
|
|||
|
|
100
|
|||
|
|
-100
|
|||
|
|
+100
|
|||
|
|
3.14159
|
|||
|
|
6.02e23
|
|||
|
|
1e-12
|
|||
|
|
''')
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
:options: +NORMALIZE_WHITESPACE
|
|||
|
|
|
|||
|
|
|
|||
|
|
# any int or real number, returned as float
|
|||
|
|
100
|
|||
|
|
[100.0]
|
|||
|
|
|
|||
|
|
-100
|
|||
|
|
[-100.0]
|
|||
|
|
|
|||
|
|
+100
|
|||
|
|
[100.0]
|
|||
|
|
|
|||
|
|
3.14159
|
|||
|
|
[3.14159]
|
|||
|
|
|
|||
|
|
6.02e23
|
|||
|
|
[6.02e+23]
|
|||
|
|
|
|||
|
|
1e-12
|
|||
|
|
[1e-12]
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
pyparsing_common.hex_integer.run_tests('''
|
|||
|
|
# hex numbers
|
|||
|
|
100
|
|||
|
|
FF
|
|||
|
|
''')
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
:options: +NORMALIZE_WHITESPACE
|
|||
|
|
|
|||
|
|
|
|||
|
|
# hex numbers
|
|||
|
|
100
|
|||
|
|
[256]
|
|||
|
|
|
|||
|
|
FF
|
|||
|
|
[255]
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
pyparsing_common.fraction.run_tests('''
|
|||
|
|
# fractions
|
|||
|
|
1/2
|
|||
|
|
-3/4
|
|||
|
|
''')
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
:options: +NORMALIZE_WHITESPACE
|
|||
|
|
|
|||
|
|
|
|||
|
|
# fractions
|
|||
|
|
1/2
|
|||
|
|
[0.5]
|
|||
|
|
|
|||
|
|
-3/4
|
|||
|
|
[-0.75]
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
pyparsing_common.mixed_integer.run_tests('''
|
|||
|
|
# mixed fractions
|
|||
|
|
1
|
|||
|
|
1/2
|
|||
|
|
-3/4
|
|||
|
|
1-3/4
|
|||
|
|
''')
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
:options: +NORMALIZE_WHITESPACE
|
|||
|
|
|
|||
|
|
|
|||
|
|
# mixed fractions
|
|||
|
|
1
|
|||
|
|
[1]
|
|||
|
|
|
|||
|
|
1/2
|
|||
|
|
[0.5]
|
|||
|
|
|
|||
|
|
-3/4
|
|||
|
|
[-0.75]
|
|||
|
|
|
|||
|
|
1-3/4
|
|||
|
|
[1.75]
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
import uuid
|
|||
|
|
pyparsing_common.uuid.set_parse_action(token_map(uuid.UUID))
|
|||
|
|
pyparsing_common.uuid.run_tests('''
|
|||
|
|
# uuid
|
|||
|
|
12345678-1234-5678-1234-567812345678
|
|||
|
|
''')
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
:options: +NORMALIZE_WHITESPACE
|
|||
|
|
|
|||
|
|
|
|||
|
|
# uuid
|
|||
|
|
12345678-1234-5678-1234-567812345678
|
|||
|
|
[UUID('12345678-1234-5678-1234-567812345678')]
|
|||
|
|
c ó> — |D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w )zK
|
|||
|
|
Parse action for converting parsed integers to Python int
|
|||
|
|
©Úint©Ú_Ú__ÚtÚtts úS/home/antigravity/intellecton/venv/lib/python3.12/site-packages/pyparsing/common.pyÚconvert_to_integerz#pyparsing_common.convert_to_integer¾ s € ð
|
|||
|
|
#$Ö$˜B”�B•Ò$Ð$ùÒ$ó …c ó> — |D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w )zL
|
|||
|
|
Parse action for converting parsed numbers to Python float
|
|||
|
|
©Úfloatr s r Úconvert_to_floatz!pyparsing_common.convert_to_floatÅ s € ð
|
|||
|
|
%&Ö&˜b”�b• Ò&Ð&ùÒ&r Úintegerc ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w ©Nr ©r r s r ú<lambda>zpyparsing_common.<lambda>Ò ó € ¨aÖ0¨œC �GÒ0€ ùÒ0r z hex integeré z[+-]?\d+zsigned integerc ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda>â r! r c ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda>ë ó € °Ö2¨"œE "�IÒ2€ ùÒ2r ú/c ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda>ñ r% r Úfractionc ó — | d | d z S )Nr éÿÿÿÿ© )r s r r zpyparsing_common.<lambda>õ s € ¨¨A©°°B±© € ó ú-z"fraction or mixed integer-fractionz[+-]?(?:\d+\.\d*|\.\d+)z real numberc ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda> r% r z@[+-]?(?:\d+(?:[eE][+-]?\d+)|(?:\d+\.\d*|\.\d+)(?:[eE][+-]?\d+)?)z$real number with scientific notationc ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda> r% r Únumberz [+-]?\d+\.?\d*(?:[eE][+-]?\d+)?Úfnumberc ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda> r% r z;(?i:[+-]?(?:(?:\d+\.?\d*(?:e[+-]?\d+)?)|nan|inf(?:inity)?))Ú
|
|||
|
|
ieee_floatc ó> — | D �cg c]
|
|||
|
|
}t |« ‘Œ c}S c c}w r r r s r r zpyparsing_common.<lambda>) r% r Ú
|
|||
|
|
identifierzQ(?:25[0-5]|2[0-4][0-9]|1?[0-9]{1,2})(?:\.(?:25[0-5]|2[0-4][0-9]|1?[0-9]{1,2})){3}z IPv4 addressz[0-9a-fA-F]{1,4}Ú hex_integerú:é zfull IPv6 address)r é z::zshort IPv6 addressc ó, — t d„ | D « « dk S )Nc 3 ó` K — | ]& }t j j |« sŒ#d –— Œ( yw)r N)r Ú
|
|||
|
|
_ipv6_partÚmatches)Ú.0r s r ú <genexpr>z,pyparsing_common.<lambda>.<locals>.<genexpr>@ s$ è ø€ ÒO˜BÔ'7×'BÑ'B×'JÑ'JÈ2Õ'N”aÑOùs ‚$.§.é )Úsum)r s r r zpyparsing_common.<lambda>@ s € ”#ÑO !ÔOÓOÐRSÑS€ r, z::ffff:zmixed IPv6 addressz IPv6 addressz:[0-9a-fA-F]{2}([:.-])[0-9a-fA-F]{2}(?:\1[0-9a-fA-F]{2}){4}z MAC addressÚfmtc ó ‡ — ˆ fd„}|S )a
|
|||
|
|
Helper to create a parse action for converting parsed date string to Python datetime.date
|
|||
|
|
|
|||
|
|
Params -
|
|||
|
|
- fmt - format to be passed to datetime.strptime (default= ``"%Y-%m-%d"``)
|
|||
|
|
|
|||
|
|
Example:
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
date_expr = pyparsing_common.iso8601_date.copy()
|
|||
|
|
date_expr.set_parse_action(pyparsing_common.convert_to_date())
|
|||
|
|
print(date_expr.parse_string("1999-12-31"))
|
|||
|
|
|
|||
|
|
prints:
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
|
|||
|
|
[datetime.date(1999, 12, 31)]
|
|||
|
|
c ó¢ •— t j |d ‰« j « S # t $ r}t | |t |« « ‚d }~ww xY w©Nr )r ÚstrptimeÚdateÚ
|
|||
|
|
ValueErrorÚParseExceptionÚstr)ÚssÚllr ÚverB s €r Úcvt_fnz0pyparsing_common.convert_to_date.<locals>.cvt_fnf sK ø€ ð
|
|||
|
|
6Ü ×(Ñ(¨¨A©°Ó4×9Ñ9Ó;Ð;øÜ ò
|
|||
|
|
6Ü$ R¨¬S°«WÓ5Ð5ûð
|
|||
|
|
6ús ƒ&* ª A³A Á Ar+ ©rB rN s ` r Úconvert_to_datez pyparsing_common.convert_to_dateO s ø€ ô. 6ð ˆ
|
|||
|
|
r, c ó ‡ — ˆ fd„}|S )a[ Helper to create a parse action for converting parsed
|
|||
|
|
datetime string to Python :class:`datetime.datetime`
|
|||
|
|
|
|||
|
|
Params -
|
|||
|
|
- fmt - format to be passed to :class:`datetime.strptime` (default= ``"%Y-%m-%dT%H:%M:%S.%f"``)
|
|||
|
|
|
|||
|
|
Example:
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
dt_expr = pyparsing_common.iso8601_datetime.copy()
|
|||
|
|
dt_expr.set_parse_action(pyparsing_common.convert_to_datetime())
|
|||
|
|
print(dt_expr.parse_string("1999-12-31T23:59:59.999"))
|
|||
|
|
|
|||
|
|
prints:
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
|
|||
|
|
[datetime.datetime(1999, 12, 31, 23, 59, 59, 999000)]
|
|||
|
|
c ó† •— t j |d ‰« S # t $ r}t | |t |« « ‚d }~ww xY wrE )r rF rH rI rJ )ÚsÚlr rM rB s €r rN z4pyparsing_common.convert_to_datetime.<locals>.cvt_fn… sB ø€ ð
|
|||
|
|
4Ü ×(Ñ(¨¨1©¨sÓ3Ð3øÜ ò
|
|||
|
|
4Ü$ Q¨¬3¨r«7Ó3Ð3ûð
|
|||
|
|
4ús ƒ œ A ¥;»A r+ rO s ` r Úconvert_to_datetimez$pyparsing_common.convert_to_datetimen s ø€ ô. 4ð ˆ
|
|||
|
|
r, z7(?P<year>\d{4})(?:-(?P<month>\d\d)(?:-(?P<day>\d\d))?)?z ISO8601 datez†(?P<year>\d{4})-(?P<month>\d\d)-(?P<day>\d\d)[T ](?P<hour>\d\d):(?P<minute>\d\d)(:(?P<second>\d\d(\.\d*)?)?)?(?P<tz>Z|[+-]\d\d:?\d\d)?zISO8601 datetimec ó — t |j j d« xs d« }t |j xs d« }t |j xs d« }t |j
|
|||
|
|
xs d« }t |j xs d« }t |j xs d« } t |||||t |« t |dz dz « « S # t $ r/} t ||d| › �« j | j « d‚d} ~ ww xY w)a Parse action to convert parsed dates or datetimes to a Python
|
|||
|
|
:class:`datetime.datetime`.
|
|||
|
|
|
|||
|
|
This parse action will use the year, month, day, etc. results
|
|||
|
|
names defined in the ISO8601 date expressions, but it can be
|
|||
|
|
used with any expression that provides one or more of these fields.
|
|||
|
|
|
|||
|
|
Omitted fields will default to fields from Jan 1, 00:00:00.
|
|||
|
|
|
|||
|
|
Invalid dates will raise a :class:`ParseException` with the
|
|||
|
|
error message indicating the invalid date fields.
|
|||
|
|
Ú0r r iè zInvalid date/time: N)r ÚyearÚlstripÚmonthÚdayÚhourÚminuter Úsecondr rH rI Úwith_tracebackÚ
|
|||
|
|
__traceback__)
|
|||
|
|
rS rT r rX rZ r[ r\ r] r^ rM s
|
|||
|
|
r Ú as_datetimez pyparsing_common.as_datetime— sï € ô �1—6‘6—=‘= Ó%Ò*¨Ó+ˆÜ�A—G‘G’L˜qÓ!ˆÜ�!—%‘%’*˜1‹oˆÜ�1—6‘6’;˜QÓ ˆÜ�Q—X‘X’] Ó#ˆÜ�q—x‘x’} 1Ó%ˆð ÜØ�e˜S $¨´°F³ ¼SÀ&È1Á*ÐPTÑATÓ=Uóð
|
|||
|
|
øô ò Ü AÐ)<¸R¸DÐ'AÓB×QÑQØ× Ñ óàð
|
|||
|
|
ûð ús Â'(C Ã DÃ*DÄDz4[0-9a-fA-F]{8}(?:-[0-9a-fA-F]{4}){3}-[0-9a-fA-F]{12}ÚUUIDrS rT Útokensc óF — t j j |d « S )a\ Parse action to remove HTML tags from web page HTML source
|
|||
|
|
|
|||
|
|
Example:
|
|||
|
|
|
|||
|
|
.. testcode::
|
|||
|
|
|
|||
|
|
# strip HTML links from normal text
|
|||
|
|
text = '<td>More info at the <a href="https://github.com/pyparsing/pyparsing/wiki">pyparsing</a> wiki page</td>'
|
|||
|
|
td, td_end = make_html_tags("TD")
|
|||
|
|
table_text = td + SkipTo(td_end).set_parse_action(
|
|||
|
|
pyparsing_common.strip_html_tags)("body") + td_end
|
|||
|
|
print(table_text.parse_string(text).body)
|
|||
|
|
|
|||
|
|
Prints:
|
|||
|
|
|
|||
|
|
.. testoutput::
|
|||
|
|
|
|||
|
|
More info at the pyparsing wiki page
|
|||
|
|
r )r Ú_html_stripperÚtransform_string)rS rT rc s r Ústrip_html_tagsz pyparsing_common.strip_html_tags s € ô* ×.Ñ.×?Ñ?ÀÀqÁ ÓJÐJr, ú,)Ú
|
|||
|
|
exclude_charsz Ú commaItemÚ )Údefaultzcomma separated listc óH — |D �cg c] }|j « ‘Œ c}S c c}w )z-Parse action to convert tokens to upper case.)Úupper©rS rT r r s r Ú
|
|||
|
|
upcase_tokensz pyparsing_common.upcase_tokensê ó € ð &'Ö'˜r�—‘•
|
|||
|
|
Ò'Ð'ùÒ'ó … c óH — |D �cg c] }|j « ‘Œ c}S c c}w )z-Parse action to convert tokens to lower case.)Úlowerro s r Údowncase_tokensz pyparsing_common.downcase_tokensï rq rr aþ (?P<url>(?:(?:(?P<scheme>https?|ftp):)?\/\/)(?:(?P<auth>\S+(?::\S*)?)@)?(?P<host>(?!(?:10|127)(?:\.\d{1,3}){3})(?!(?:169\.254|192\.168)(?:\.\d{1,3}){2})(?!172\.(?:1[6-9]|2\d|3[0-1])(?:\.\d{1,3}){2})(?:[1-9]\d?|1\d\d|2[01]\d|22[0-3])(?:\.(?:1?\d{1,2}|2[0-4]\d|25[0-5])){2}(?:\.(?:[1-9]\d?|1\d\d|2[0-4]\d|25[0-4]))|(?:(?:[a-z0-9\u00a1-\uffff][a-z0-9\u00a1-\uffff_-]{0,62})?[a-z0-9\u00a1-\uffff]\.)+(?:[a-z\u00a1-\uffff]{2,}\.?))(:(?P<port>\d{2,5}))?(?P<path>\/[^?# ]*)?(\?(?P<query>[^#]*))?(#(?P<fragment>\S*))?)ÚurlÚconvertToIntegerÚconvertToFloatÚ
|
|||
|
|
convertToDateÚconvertToDatetimeÚ
|
|||
|
|
stripHTMLTagsÚ upcaseTokensÚdowncaseTokensN)z%Y-%m-%d)z%Y-%m-%dT%H:%M:%S.%f)OÚ__name__Ú
|
|||
|
|
__module__Ú __qualname__Ú__doc__Ú staticmethodr r ÚWordÚnumsÚset_nameÚset_parse_actionÚPY_310_OR_LATERr ÚhexnumsÚ token_mapr r6 ÚRegexÚsigned_integerr( Úadd_parse_actionÚOptÚsuppressÚ
|
|||
|
|
mixed_integerrA ÚrealÚsci_realÚ
|
|||
|
|
streamliner0 r1 r3 Ú
|
|||
|
|
identcharsÚidentbodycharsr5 Ú ipv4_addressr< Ú_full_ipv6_addressÚ_short_ipv6_addressÚ
|
|||
|
|
add_conditionÚ_mixed_ipv6_addressÚCombineÚ ipv6_addressÚ mac_addressrJ rP rU Ú iso8601_dateÚiso8601_datetimera Úiso8601_date_validatedÚiso8601_datetime_validatedÚuuidr r re Ú ParseResultsrg Ú OneOrMoreÚLiteralÚLineEndÚ
|
|||
|
|
printablesÚWhiteÚ
|
|||
|
|
FollowedByÚ
|