gE d Z ddlZddlZddlmZ ddlmZmZmZm Z ddl
m Z m Z G d d Z
G d d e
Z ej d
Zdd Z G d d
e
ZddZd Zd Zd Zd Zd Zd Zd Zedk( r e yy)zT
Module for reading, writing and manipulating
Toolbox databases and settings files.
N)StringIO)Element ElementTree
SubElement TreeBuilder) PathPointerfindc B e Zd ZdZd dZd Zd Zd Z d
dZd Z y) StandardFormatzT
Class for reading and processing standard format marker files and strings.
Nc : || _ || j | y y N) _encodingopen)selffilenameencodings A/var/www/openai/venv/lib/python3.12/site-packages/nltk/toolbox.py__init__zStandardFormat.__init__ s ! IIh c t |t r!|j | j | _ yt j |d| j | _ y)z
Open a standard format marker file for sequential reading.
:param sfm_file: name of the standard format marker input file
:type sfm_file: str
rN)
isinstancer r r _filecodecs)r sfm_files r r zStandardFormat.open s: h ,! t~~6DJ XsDNNCDJr c $ t | | _ y)z
Open a standard format marker string for sequential reading.
:param s: string to parse as a standard format marker input file
:type s: str
N)r r )r ss r open_stringzStandardFormat.open_string+ s a[
r c # t K d}d}t j |dz }t j |dz }t | j } t | }t j || }|j \ }} | g}
d| _ |D ]n }| xj dz
c_ t j || }|j \ } } | r||j |
f | }| g}
^|
j | p | xj dz
c_ ||j |
f y# t
$ r Y yw xY ww)z
Return an iterator that returns the next field in a (marker, value)
tuple. Linebreaks and trailing white space are preserved except
for the final newline in each field.
:rtype: iter(tuple(str, str))
z^%s(?:\\(\S+)\s*)?(.*)$u (?:)? Nr ) recompileiterr next
StopIterationmatchgroupsline_numjoinappend) r join_string line_regexpfirst_line_patline_pat file_iterlinemobjmkr
line_value value_linesline_mkrs r
raw_fieldszStandardFormat.raw_fields4 s* 0 K2E$EF::kB ./ $ ?D xx- ++-Z!l
D MMQ M88Hd+D#';;= HjK , ,[ 9:: )l "":.
K$$[122% s+ A
D8
D) CD8) D52D84D55D8c # K |
| t d t j d }| j D ]1 \ }}|r|j d| }|r|j }||f 3 yw)a
Return an iterator that returns the next field in a ``(marker, value)``
tuple, where ``marker`` and ``value`` are unicode strings if an ``encoding``
was specified in the ``fields()`` method. Otherwise they are non-unicode strings.
:param strip: strip trailing whitespace from the last line of each field
:type strip: bool
:param unwrap: Convert newlines in a field to spaces.
:type unwrap: bool
:param encoding: Name of an encoding to use. If it is specified then
the ``fields()`` method returns unicode strings rather than non
unicode strings.
:type encoding: str or None
:param errors: Error handling scheme for codec. Same as the ``decode()``
builtin string method.
:type errors: str
:param unicode_fields: Set of marker names whose values are UTF-8 encoded.
Ignored if encoding is None. If the whole file is UTF-8 encoded set
``encoding='utf8'`` and leave ``unicode_fields`` with its default
value of None.
:type unicode_fields: sequence
:rtype: iter(tuple(str, str))
Nz'unicode_fields is set but not encoding.z\n+ )
ValueErrorr# r$ r8 subrstrip) r stripunwrapr errorsunicode_fields
unwrap_patr4 vals r fieldszStandardFormat.fields[ sm > : FG GZZ'
)HC nnS#.jjl* *s A*A,c \ | j j | `y# t $ r Y yw xY w)z@Close a previously opened standard format marker file or string.N)r closer* AttributeError)r s r rF zStandardFormat.close s.
s ++)NN)TTNstrictN)
__name__
__module__ __qualname____doc__r r r r8 rD rF r r r r s9
D!%3R
' Rr r c * e Zd ZddZddZd ZddZy) ToolboxDataNc R |r | j dd|i|S | j di |S )NgrammarrM ) _chunk_parse
_record_parse)r rQ kwargss r parsezToolboxData.parse s6 $4$$?W?? ?%4%%// /r c " t }|j di |j di d} | j di |D ] \ }}| |s
|d dk7 r|}||k( r9|r|j d n|j d d}|j di |j |i |j | |j | |r|j d n|j d |j d |j S ) a
Returns an element tree structure corresponding to a toolbox data file with
all markers at the same level.
Thus the following Toolbox database::
\_sh v3.0 400 Rotokas Dictionary
\_DateStampHasFourDigitYear
\lx kaa
\ps V.A
\ge gag
\gp nek i pas
\lx kaa
\ps V.B
\ge strangle
\gp pasim nek
after parsing will end up with the same structure (ignoring the extra
whitespace) as the following XML fragment after being parsed by
ElementTree::
<_sh>v3.0 400 Rotokas Dictionary
<_DateStampHasFourDigitYear/>
kaa
V.A
gag
nek i pas
kaa
V.B
strangle
pasim nek
:param key: Name of key marker at the start of each record. If set to
None (the default value) the first marker that doesn't begin with
an underscore is assumed to be the key.
:type key: str
:param kwargs: Keyword arguments passed to ``StandardFormat.fields()``
:type kwargs: dict
:rtype: ElementTree._ElementInterface
:return: contents of toolbox data divided into header and records
toolbox_dataheaderFr _recordTrM )r startrD enddatarF )r keyrT builder
in_recordsr4 values r rS zToolboxData._record_parse s h -
nb)
h#
%$++//JC{:#a&C-cz KK)KK)!%J
h+ MM#r " LL KK 0 KK ! KK ! N#}} r c ddl m} t |j }|D ]G }t || r!|j | j
| 0|\ }}t || }||_ I |S )Nr Tree) nltk.treerd r labelr r, _tree2etreer text)r parentrd rootchildrh tages r rg zToolboxData._tree2etree sc "v||~&E%& D , ,U 34 ! c tS)
r c ddl m} ddlm} |j ||| } | j
di |}t
d } |j d }
| j |
|j d D ]W } |j | D cg c] } | j | j f c} }
| j | j |
Y | S c c} w ) aw
Returns an element tree structure corresponding to a toolbox data file
parsed according to the chunk grammar.
:type grammar: str
:param grammar: Contains the chunking rules used to parse the
database. See ``chunk.RegExp`` for documentation.
:type root_label: str
:param root_label: The node value that should be used for the
top node of the chunk structure.
:type trace: int
:param trace: The level of tracing that should be used when
parsing a text. ``0`` will generate no tracing output;
``1`` will generate normal tracing output; and ``2`` or
higher will generate verbose tracing output.
:type kwargs: dict
:param kwargs: Keyword arguments passed to ``toolbox.StandardFormat.fields()``
:rtype: ElementTree._ElementInterface
r )chunkrc )
root_labeltracerW rX rZ rM )
nltkro re rd RegexpParserrU r r r, findallrh rl rg )r rQ rp rq rT ro rd cpdbtb_etreerX rZ elemparseds r rR zToolboxData._chunk_parse s ( "
Je
L
TZZ
!&
!>*" jj*FXX H 488 4 HIF OOD , ,V 4 5 + Is C
r
)NrZ r )rI rJ rK rU rS rg rR rM r r rO rO s 0 J X r rO z\Sc | j dk( r t d }|j | |} | j dk7 r t d |
| t d g }| D ] }|j d |D ] }|j }|j } |v|||v rd}
n|}
t j t | r)|j d| d | dj |
| k|j d| | dj |
| t j t | r|j d| d | d |j d| | d d
j |d d S ) a@
Return a string with a standard format representation of the toolbox
data in tree (tree can be a toolbox database or a single record).
:param tree: flat representation of toolbox data (whole database or single record)
:type tree: ElementTree._ElementInterface
:param encoding: Name of an encoding to use.
:type encoding: str
:param errors: Error handling scheme for codec. Same as the ``encode()``
builtin string method.
:type errors: str
:param unicode_fields:
:type unicode_fields: dict(str) or set(str)
:rtype: str
rZ rW z$not a toolbox_data element structureNz?if encoding is not specified then neither should unicode_fieldsr utf8\r: r! r" )
rl r r, r; rh r# search _is_valueencoder+ ) treer r@ rA rj lrecfieldr4 ra cur_encodings r
to_sfm_stringr sg xx8~& D xx>!?@@N 6 M
A E))CJJE#!-# 2G#)L#+L99Y.HH3%qr 2 : :< PQHH3%wb 1 9 9, OP99Y.HHr#awb 12HHr#ugR 01! & 771QR5>r c * e Zd ZdZ fdZddZ xZS )ToolboxSettingsz0This class is the base class for settings files.c " t | y r
)superr )r __class__s r r zToolboxSettings.__init__G s
r c z t } | j d||d|D ] \ }}|d }|dv r|dd }nd}|dk( r$|j |i |j | @|dk( r|j | W|j |i |j | |j | |j S ) a
Return the contents of toolbox settings file with a nested structure.
:param encoding: encoding used by settings file
:type encoding: str
:param errors: Error handling scheme for codec. Same as ``decode()`` builtin method.
:type errors: str
:param kwargs: Keyword arguments passed to ``StandardFormat.fields()``
:type kwargs: dict
:rtype: ElementTree._ElementInterface
)r r@ r )+-r" Nr r rM )r rD r[ r] r\ rF )r r r@ rT r_ r4 ra blocks r rU zToolboxSettings.parseJ s -%$++QxQ&QJC FE
"!"g |
c2& U## C
c2& U# C # R$ }} r )NrH )rI rJ rK rL r rU
__classcell__)r s @r r r D s : r r c t t }t | j |||| dj | S )N)r r@ rA r! )list_to_settings_stringgetrootr+ )r r r@ rA r s r to_settings_stringr l s6 A % 771:r c ^ | j }| j }t | dk( r0|r|j d| d| d y |j d|z y |r|j d| d| d n|j d|z | D ] }t ||fi | |j d|z y ) Nr r| r: r z\%s
z\+z\+%s
z\-%s
)rl rh lenr, r )noder rT rl rh ns r r r y s
((C 99D
4yA~
HHr#avR( )
HHX^ $
HHs3%qb) *
HHY_ %A 1 / / S !
r c t }| D ]9 }t | |j st | dkD s)|j | ; || dd y)z
Remove all elements and subelements with no text and no child elements.
:param elem: toolbox data in an elementtree structure
:type elem: ElementTree._ElementInterface
r N)r
remove_blanksrh r r, )rx outrk s r r r sE &Ce ::Ua JJu DGr c |j | j g D ] }| j | t | | " | D ] }t || y)a(
Add blank elements and subelements specified in default_fields.
:param elem: toolbox data in an elementtree structure
:type elem: ElementTree._ElementInterface
:param default_fields: fields to add to each type of element and subelement
:type default_fields: dict(tuple)
N)getrl r r add_default_fields)rx default_fieldsr rk s r r r sN ##DHHb1 99U # tU # 2 5.1 r c t }|j D ], \ }}t x||< }t | D ]
\ }}|||< . t | | y)a-
Sort the elements and subelements in order specified in field_orders.
:param elem: toolbox data in an elementtree structure
:type elem: ElementTree._ElementInterface
:param field_orders: order of fields for each type of element and subelement
:type field_orders: dict(tuple)
N)dictitems enumerate _sort_fields)rx field_orders order_dictsr order order_keyisubfields r sort_fieldsr s\ &K$**, u)-/ EY$U+KAx"#Ih , - {#r c || j t fdt | D }|D cg c] \ }}| c}}| dd | D ] }t | st || yc c}}w # t $ r Y 1w xY w)zsort the children of elemc 3 d K | ]' \ }}j |j d |f|f ) yw)g eAN)r rl ).0r rk r s r z _sort_fields.. s2
DS5eii 3'
+U 3Os -0N)rl sortedr KeyErrorr r )rx orders_dictstmpr^ rk r s @r r r s 0 TXX&
DMdO
,//3ZS%53/Q u: - 0
s A- A'- A98A9c || j }|| j }d}| D ] }|j }|A|j |k7 r2||v r]|[|j d }|j xs ddz |_ n/||v r+|j d }|j xs ddz |_ t | r
t ||| |} y# t
$ r# | D ] }t | st ||| Y yw xY w)a1
Add blank lines before all elements and subelements specified in blank_before.
:param elem: toolbox data in an elementtree structure
:type elem: ElementTree._ElementInterface
:param blank_before: elements and subelements to add blank lines before
:type blank_before: dict(tuple)
Nr! r )rl getiteratorrh r add_blank_linesr ) r
blanks_beforeblanks_betweenbeforebetween last_elemrx rl rm s r r r s txx( * D((C$ #)=&=Y%:!--/3A ffld 2AF'>!--/3A ffld 2AF4y m^D I
ED4y m^D Es C C. C.-C.c ddl m} t d }t | j }t d t |d d j t |d d j t d | |j d d D ]" }t |j |j $ t d | |j d
d D ] }t |j t }t d }|j | |j d d
}t |j d j t | }t t | j d y )Nr )islicezcorpora/toolbox/rotokas.dicz first field in fourth record: z
fields in sequential order:rZ
z
lx fields:z record/lxz corpora/toolbox/MDF/MDF_AltH.typFcp1252)r? r z$expset/expMDF/rtfPageSetup/paperSizer{ ) itertoolsr r rO rU printrl rh rt r r r r r )r file_pathlexiconr settingsr
settings_trees r demor s- 23I)$**,G
)* '!*Q-
'!*Q-
)* X.3
eii$ 4
. 4b9
ejj : H78I MM) >>> :D $)):
;
@
@A %M
]
+
2
26
:;r __main__)NrH N) rL r r# ior xml.etree.ElementTreer r r r nltk.datar r r rO r$ r~ r r r r r r r r r r rI rM r r r s
O O 'u up. D
BJJu /d% n % P
(
2 $"
. B<: zF r