Das ist rausgekommen:
Code: Alles auswählen
#!/usr/bin/env python
"""
Pygments Tkinter example
copyleft 2014 by Jens Diemer
licensed under GNU GPL v3
"""
import os
try:
# Python 3
import tkinter
from tkinter import font
from tkinter import scrolledtext
except ImportError:
# Python 2
import Tkinter as tkinter
import tkFont as font
import ScrolledText as scrolledtext
from pygments.lexers.python import PythonLexer
from pygments.styles import get_style_by_name
class TkTest(object):
def __init__(self, lexer):
self.lexer = lexer
self.root = tkinter.Tk()
self.root.title("%s (%s)" % (lexer.name, os.path.basename(__file__)))
self.text = scrolledtext.ScrolledText(
master=self.root, height=30, width=120
)
self.text.pack(side=tkinter.LEFT, fill=tkinter.Y)
self.root.bind("<Key>", self.event_key)
self.root.bind('<Control-KeyPress-a>', self.event_select_all)
self.root.bind('<Control-KeyPress-x>', self.event_cut)
self.root.bind('<Control-KeyPress-c>', self.event_copy)
self.root.bind('<Control-KeyPress-v>', self.event_paste)
self.root.update()
self.create_tags()
def event_select_all(self, event=None):
self.text.tag_add(tkinter.SEL, "1.0", tkinter.END)
self.text.mark_set(tkinter.INSERT, "1.0")
self.text.see(tkinter.INSERT)
return "break"
def event_cut(self, event=None):
if self.text.tag_ranges(tkinter.SEL):
self.event_copy()
self.text.delete(tkinter.SEL_FIRST, tkinter.SEL_LAST)
self.recolorize()
return "break"
def event_copy(self, event=None):
if self.text.tag_ranges(tkinter.SEL):
text = self.text.get(tkinter.SEL_FIRST, tkinter.SEL_LAST)
self.clipboard_clear()
self.clipboard_append(text)
return "break"
def paste(self, text):
if text:
self.text.insert(tkinter.INSERT, text)
self.text.tag_remove(tkinter.SEL, '1.0', tkinter.END)
self.text.see(tkinter.INSERT)
self.recolorize()
def event_paste(self, event=None):
text = self.text.selection_get(selection='CLIPBOARD')
self.paste(text)
return "break"
def event_key(self, event):
keycode = event.keycode
char = event.char
print("\tkeycode %s - char %s" % (keycode, repr(char)))
self.recolorize()
def mainloop(self):
self.root.mainloop()
# ---------------------------------------------------------------------------------------
def create_tags(self):
bold_font = font.Font(self.text, self.text.cget("font"))
bold_font.configure(weight=font.BOLD)
italic_font = font.Font(self.text, self.text.cget("font"))
italic_font.configure(slant=font.ITALIC)
bold_italic_font = font.Font(self.text, self.text.cget("font"))
bold_italic_font.configure(weight=font.BOLD, slant=font.ITALIC)
style = get_style_by_name('default')
for ttype, ndef in style:
# print(ttype, ndef)
tag_font = None
if ndef['bold'] and ndef['italic']:
tag_font = bold_italic_font
elif ndef['bold']:
tag_font = bold_font
elif ndef['italic']:
tag_font = italic_font
if ndef['color']:
foreground = "#%s" % ndef['color']
else:
foreground = None
self.text.tag_configure(str(ttype), foreground=foreground, font=tag_font)
def recolorize(self):
print("recolorize")
code = self.text.get("1.0", "end-1c")
tokensource = self.lexer.get_tokens(code)
start_line=1
start_index = 0
end_line=1
end_index = 0
for ttype, value in tokensource:
if "\n" in value:
end_line += value.count("\n")
end_index = len(value.rsplit("\n",1)[1])
else:
end_index += len(value)
if value not in (" ", "\n"):
index1 = "%s.%s" % (start_line, start_index)
index2 = "%s.%s" % (end_line, end_index)
for tagname in self.text.tag_names(index1): # FIXME
self.text.tag_remove(tagname, index1, index2)
# print(ttype, repr(value), index1, index2)
self.text.tag_add(str(ttype), index1, index2)
start_line = end_line
start_index = end_index
def removecolors(self):
for tag in self.tagdefs:
self.text.tag_remove(tag, "1.0", "end")
if __name__ == "__main__":
tk_win = TkTest(lexer = PythonLexer())
with open(__file__, "r") as f:
tk_win.paste(f.read())
tk_win.mainloop()
Optimal finde ich das ganze allerdings nicht:
1. es wird immer alles neu mit tags versehen
2. Bei .tag_add() bestehende tags nicht überschreibt, wird vorher mit tag_remove() alle möglichen Tags entfernt
Also irgendwie nicht besonders effektiv.
Wobei zu 1.: Man kann auch nicht einfach sagen, wir aktualisieren nur immer eine Zeile: Das würde IMHO bei Block-Referenzen nicht greifen (Also sowas wie Block-Kommentare in mehreren Zeilen)
Verbesserungsvorschläge?