Skip to content

GitLab

D
DHParser

In January 2021 we will introduce a 10 GB quota for project repositories. Higher limits for individual projects will be available on request. Please see https://doku.lrz.de/display/PUBLIC/GitLab for more information.

Commit fc3aa0d1 authored by di68kap's avatar di68kap
Browse files
Options

- MLW ergänzt

parent 56729685
Hide whitespace changes
Inline Side-by-side
Showing with 77 additions and 45 deletions
DHParser/syntaxtree.py
View file @ fc3aa0d1
......@@ -486,8 +486,8 @@ class Node(collections.abc.Sized):
# s += " '(pos %i)" % node.pos
if src:
txt += " '(pos %i " % node.pos # + " %i %i)" % line_col(src, node.pos)
if node.error_flag:
txt += " HAS ERRORS"
# if node.error_flag: # just for debugging error collecting
# txt += " HAS ERRORS"
if node.errors:
txt += " '(err '(%s))" % ' '.join(str(err).replace('"', r'\"')
for err in node.errors)
......@@ -556,8 +556,8 @@ class Node(collections.abc.Sized):
"""
Finds nodes in the tree that match a specific criterion.
``find`` is a generator that yields all nodes for which the
given ``match_function`` evaluates to True. The tree is
`find` is a generator that yields all nodes for which the
given `match_function` evaluates to True. The tree is
traversed pre-order.
Args:
......@@ -596,7 +596,8 @@ class Node(collections.abc.Sized):
def mock_syntax_tree(sxpr):
"""
Generates a tree of nodes from an S-expression.
Generates a tree of nodes from an S-expression. The main purpose of this is
to generate test data.
Example:
>>> mock_syntax_tree("(a (b c))").as_sxpr()
......
examples/MLW/Beispiele/imperium/imperium.mlw
View file @ fc3aa0d1
......@@ -11,7 +11,8 @@ SCHREIBWEISE
em-: Chron. Fred. 2,35sqq. capit. p. 43. 2,36 p. 60,10.
ym-: Chart. Sangall. A 194.
impir-: v. ibi.
STRUKTUR
form. sing.:
gen.:
-ri: v. ibi. adde Annal. Plac. a. 1266 p. 516,21.
......@@ -19,6 +20,7 @@ SCHREIBWEISE
abl.:
-um: Chron. Fred. 2,15. 2,35sqq. capit. p. 43.
VERWECHSELBAR
confunditur c.:
imperitus: v. ibi.
......
examples/MLW/MLW.ebnf
View file @ fc3aa0d1
......@@ -32,6 +32,8 @@ Lemma = [< klassisch | gesichert >] LemmaWort
klassisch = "*"
gesichert = "$" # TODO: Noch fragen: Welches Zeichen?
LemmaWort = LAT_WORT
LemmaVarianten = [LZ]
{ LemmaVariante §TR }+
[Zusatz]
......@@ -76,7 +78,7 @@ Etymologie = FREITEXT
#### ARTIKEL-KOPF ############################################################
ArtikelKopf = < SchreibweisenPosition | StrukturPosition | GebrauchPosition
| MetrikPosition | VerwechselungPosition >
| MetrikPosition | VerwechselungsPosition >
## Schreibweisen-Position ##
......@@ -84,35 +86,63 @@ ArtikelKopf = < SchreibweisenPosition | StrukturPosition | GebrauchPositio
SchreibweisenPosition = "SCHREIBWEISE" [LZ] { SWKategorie }+
SWKategorie = SWTyp DPP [LZ] §{SWUnterkategorie | SWVariante { ABS SWVariante }}+ [LZ]
SWUnterkategorie = SWUnterTyp DPP [LZ] §SWVariante { ABS SWVariante }+ [LZ]
SWTyp = scriptfat | scriptform | script | form | OFFEN
SWVariante = !KATEGORIENZEILE Schreibweise DPP Beleg
Schreibweise = ZEICHENFOLGE
SWKategorie = SWTyp DPP [LZ] §(Varianten | { SWKategorie }+) [LZ]
SWTyp = scriptfat | scriptform | script | form | gen | abl | OFFEN
scriptfat = "script." "fat-"
scriptform = "script. " "form"
script = "srcipt."
form = "form"
gen = "gen."
abl = "abl."
OFFEN = FREITEXT
#### STRUKTUR-POSITION #######################################################
StrukturPosition = "STRUKTUR" [LZ] { STVariante }+
StrukturPosition = "STRUKTUR" [LZ] { STKategorie }+
STKategorie = STTyp DPP [LZ] §(Varianten | { STKategorie }+) [LZ]
STTyp = "form." ZEICHENFOLGE
#### GEBRAUCH-POSITION #######################################################
GebrauchPosition = "GEBRAUCH"
#### METRIK-POSITION #########################################################
STVariante = "TODO"
MetrikPosition = "METRIK"
#### VERWECHSLUNGS-POSITION ##################################################
VerwechslungsPosition = "VERWECHSELBAR" [LZ] [Confunditur]
VerwechselungsVariante { ABS VerwechselungsVariante }+ [LZ]
Confunditur = "confunditur c." DPP [LZ]
VerwechselungsVariante = !KATEGORIENZEILE Verwechselung DPP Beleg
Verwechselung = ZEICHENFOLGE
#### ARTIKELKOPF POSITIONEN VARIANTEN ########################################
Varianten = Variante { ABS Variante }
Variante = !KATEGORIENZEILE Gegenstand DPP Beleg
Gegenstand = ZEICHENFOLGE
#### BEDEUTUNGS-POSITION #####################################################
##############################################################################
BedeutungsPosition = { "BEDEUTUNG" [LZ] §Bedeutung }+
Bedeutung = (Interpretamente | Bedeutungskategorie) [Belege]
Bedeutungskategorie = /(?:(?![A-ZÄÖÜ][A-ZÄÖÜ]).)+/~ [LZ]
Interpretamente = LateinischeBedeutung [LZ] §DeutscheBedeutung [LZ]
LateinischeBedeutung = SW_LAT /(?:(?![A-ZÄÖÜ][A-ZÄÖÜ]).)+/~
DeutscheBedeutung = SW_DEU /(?:(?![A-ZÄÖÜ][A-ZÄÖÜ]).)+/~
LateinischeBedeutung = LAT /(?:(?![A-ZÄÖÜ][A-ZÄÖÜ]).)+/~
DeutscheBedeutung = DEU /(?:(?![A-ZÄÖÜ][A-ZÄÖÜ]).)+/~
Belege = "BELEGE" [LZ] (EinBeleg | { "*" EinBeleg }) ABS
EinBeleg = { !([LZ] "*" | SCHLUESSELWORT) /\s*[^\n]*/~ [ZW] }+ [Zusatz]
......@@ -137,9 +167,9 @@ Name = { NAME | NAMENS_ABKÜRZUNG }+
#### Schlüsselwörter #########################################################
SW_LAT = "LATEINISCH" | "LAT"
SW_DEU = "DEUTSCH" | "DEU"
SW_GRIECH = "GRIECHISCH" | "GRIECH" | "GRIE" | "GRI"
LAT = "LATEINISCH" | "LAT"
DEU = "DEUTSCH" | "DEU"
GRI = "GRIECHISCH" | "GRIECH" | "GRIE" | "GRI"
SCHLUESSELWORT = { //~ /\n/ }+ !ROEMISCHE_ZAHL /[A-ZÄÖÜ]{3,}\s+/
......@@ -162,6 +192,11 @@ Verweis = "->" ZielName
VerweisZiel = "{" ZielName "}"
ZielName = BUCHSTABENFOLGE
Autor = FREITEXT
Werk = FREITEXT
Stelle = FREITEXT
Datierung = FREITEXT
Edition = FREITEXT
#### GENERISCHE UND ATOMARE AUSDRÜCKE ########################################
......@@ -177,9 +212,12 @@ LAT_WORT_TEIL = /[a-z]+/
GROSSSCHRIFT = /[A-ZÄÖÜ]+/~
ZAHL = /\d+/~
ROEMISCHE_ZAHL = /(?=[MDCLXVI])M*(C[MD]|D?C*)(X[CL]|L?X*)(I[XV]|V?I*)/~
SATZZEICHEN = /(?:,(?!,))|(?:;(?!;))|(?::(?!:))|[.()-]+/~ # div. Satzzeichen, aber keine doppelten ,, ;; oder ::
FREITEXT = { DEU_WORT | ZAHL | ROEMISCHE_ZAHL | SATZZEICHEN | /s*/ }+
SATZZEICHEN = /(?:,(?!,))|(?:;(?!;))|(?::(?!:))|[.()\-]+/~ # div. Satzzeichen, aber keine doppelten ,, ;; oder ::
TEXTELEMENT = DEU_WORT | ZAHL | ROEMISCHE_ZAHL
FREITEXT = { TEXTELEMENT | /[.()\-\s]+/ | /,(?!,)\s*/ }+
ERW_FREITEXT = { TEXTELEMENT | SATZZEICHEN | /\s+/ }+
BUCHSTABENFOLGE = /\w+/~
ZEICHENFOLGE = /[\w()-]+/~
......
examples/MLW/MLWCompiler.py
View file @ fc3aa0d1
......@@ -16,7 +16,7 @@ except ImportError:
import re
from DHParser import logging, is_filename, load_if_file, \
Grammar, Compiler, nil_preprocessor, \
Lookbehind, Lookahead, Alternative, Pop, Required, Token, Synonym, \
Lookbehind, Lookahead, Alternative, Pop, Token, Synonym, \
Option, NegativeLookbehind, OneOrMore, RegExp, Retrieve, Series, RE, Capture, \
ZeroOrMore, Forward, NegativeLookahead, mixin_comment, compile_source, \
last_value, counterpart, accumulate, PreprocessorFunc, \
......@@ -26,7 +26,6 @@ from DHParser import logging, is_filename, load_if_file, \
remove_expendables, remove_empty, remove_tokens, flatten, is_whitespace, \
is_empty, is_expendable, collapse, replace_content, remove_parser, remove_content, remove_brackets, replace_parser, \
keep_children, is_one_of, has_content, apply_if, remove_first, remove_last
from DHParser.base import WHITESPACE_PTYPE, TOKEN_PTYPE
#######################################################################
......
examples/MLW/MLW_ebnf_ERRORS.txt
View file @ fc3aa0d1
line: 31, column: 49, Error: Missing definition for symbol 'LemmaWort'
line: 70, column: 22, Error: Missing definition for symbol 'LAT'
line: 70, column: 28, Error: Missing definition for symbol 'GRI'
line: 78, column: 66, Error: Missing definition for symbol 'GebrauchPosition'
line: 79, column: 23, Error: Missing definition for symbol 'MetrikPosition'
line: 79, column: 40, Error: Missing definition for symbol 'VerwechselungPosition'
line: 88, column: 20, Error: Missing definition for symbol 'SWUnterTyp'
line: 89, column: 61, Error: Missing definition for symbol 'OFFEN'
line: 142, column: 1, Warning: Rule "SW_GRIECH" is not connected to parser root "Artikel" !
line: 158, column: 20, Error: Missing definition for symbol 'Autor'
line: 158, column: 30, Error: Missing definition for symbol 'Werk'
line: 158, column: 39, Error: Missing definition for symbol 'Stelle'
line: 158, column: 51, Error: Missing definition for symbol 'Datierung'
line: 158, column: 67, Error: Missing definition for symbol 'Edition'
line: 162, column: 1, Warning: Rule "VerweisZiel" is not connected to parser root "Artikel" !
line: 173, column: 1, Warning: Rule "DEU_GROSS" is not connected to parser root "Artikel" !
line: 174, column: 1, Warning: Rule "DEU_KLEIN" is not connected to parser root "Artikel" !
line: 175, column: 1, Warning: Rule "LAT_WORT" is not connected to parser root "Artikel" !
line: 177, column: 1, Warning: Rule "GROSSSCHRIFT" is not connected to parser root "Artikel" !
line: 199, column: 1, Warning: Rule "RZS" is not connected to parser root "Artikel" !
line: 206, column: 1, Warning: Rule "NIEMALS" is not connected to parser root "Artikel" !
line: 208, column: 1, Warning: Rule "DUMMY" is not connected to parser root "Artikel" !
line: 81, column: 40, Error: Missing definition for symbol 'VerwechselungsPosition'
line: 126, column: 1, Warning: Rule "VerwechslungsPosition" is not connected to parser root "Artikel" !
line: 128, column: 1, Warning: Rule "Confunditur" is not connected to parser root "Artikel" !
line: 129, column: 1, Warning: Rule "VerwechselungsVariante" is not connected to parser root "Artikel" !
line: 130, column: 1, Warning: Rule "Verwechselung" is not connected to parser root "Artikel" !
line: 188, column: 1, Warning: Rule "VerweisZiel" is not connected to parser root "Artikel" !
line: 204, column: 1, Warning: Rule "DEU_GROSS" is not connected to parser root "Artikel" !
line: 205, column: 1, Warning: Rule "DEU_KLEIN" is not connected to parser root "Artikel" !
line: 208, column: 1, Warning: Rule "GROSSSCHRIFT" is not connected to parser root "Artikel" !
line: 212, column: 1, Warning: Rule "SATZZEICHEN" is not connected to parser root "Artikel" !
line: 216, column: 1, Warning: Rule "ERW_FREITEXT" is not connected to parser root "Artikel" !
line: 233, column: 1, Warning: Rule "RZS" is not connected to parser root "Artikel" !
line: 240, column: 1, Warning: Rule "NIEMALS" is not connected to parser root "Artikel" !
line: 242, column: 1, Warning: Rule "DUMMY" is not connected to parser root "Artikel" !
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment