mirror of
https://github.com/chubin/cheat.sh.git
synced 2024-12-14 09:42:25 +03:00
569 lines
18 KiB
Python
569 lines
18 KiB
Python
import gevent
|
|
from gevent.wsgi import WSGIServer
|
|
from gevent.queue import Queue
|
|
from gevent.monkey import patch_all
|
|
from gevent.subprocess import Popen, PIPE, STDOUT
|
|
patch_all()
|
|
|
|
import sys
|
|
import os
|
|
import glob
|
|
import re
|
|
import random
|
|
import string
|
|
|
|
from fuzzywuzzy import process, fuzz
|
|
|
|
import redis
|
|
import colored
|
|
|
|
from pygments import highlight as pygments_highlight
|
|
from pygments.lexers import BashLexer, GoLexer, ScalaLexer, RustLexer, PythonLexer, PhpLexer
|
|
from pygments.formatters import TerminalFormatter, Terminal256Formatter
|
|
|
|
from pygments.styles import get_all_styles
|
|
COLOR_STYLES = sorted(list(get_all_styles()))
|
|
|
|
MYDIR = os.path.abspath(os.path.dirname(os.path.dirname('__file__')))
|
|
sys.path.append("%s/lib/" % MYDIR)
|
|
from globals import error, ANSI2HTML, \
|
|
PATH_TLDR_PAGES, PATH_CHEAT_PAGES, \
|
|
PATH_CHEAT_SHEETS, PATH_CHEAT_SHEETS_SPOOL
|
|
|
|
from buttons import TWITTER_BUTTON, GITHUB_BUTTON, GITHUB_BUTTON_2, GITHUB_BUTTON_FOOTER
|
|
|
|
from adapter_learnxiny import get_learnxiny, get_learnxiny_list, is_valid_learnxy
|
|
|
|
# globals
|
|
INTERNAL_TOPICS = [":firstpage", ':post', ':bash_completion', ':help', ':styles', ':styles-demo']
|
|
LEXER = {
|
|
"go" : GoLexer,
|
|
"scala" : ScalaLexer,
|
|
"rust" : RustLexer,
|
|
"python": PythonLexer,
|
|
"php" : PhpLexer,
|
|
}
|
|
REDIS = redis.StrictRedis(host='localhost', port=6379, db=0)
|
|
|
|
def update_tldr_topics():
|
|
answer = []
|
|
for topic in glob.glob(PATH_TLDR_PAGES):
|
|
_, filename = os.path.split(topic)
|
|
if filename.endswith('.md'):
|
|
answer.append(filename[:-3])
|
|
return answer
|
|
TLDR_TOPICS = update_tldr_topics()
|
|
|
|
def update_cheat_topics():
|
|
answer = []
|
|
for topic in glob.glob(PATH_CHEAT_PAGES):
|
|
_, filename = os.path.split(topic)
|
|
answer.append(filename)
|
|
return answer
|
|
CHEAT_TOPICS = update_cheat_topics()
|
|
|
|
def update_cheat_sheets_topics():
|
|
answer = []
|
|
answer_dirs = []
|
|
|
|
for topic in glob.glob(PATH_CHEAT_SHEETS + "*/*"):
|
|
dirname, filename = os.path.split(topic)
|
|
dirname = os.path.basename(dirname)
|
|
if dirname.startswith('_'):
|
|
dirname = dirname[1:]
|
|
answer.append("%s/%s" % (dirname, filename))
|
|
|
|
for topic in glob.glob(PATH_CHEAT_SHEETS + "*"):
|
|
_, filename = os.path.split(topic)
|
|
if os.path.isdir(topic):
|
|
if filename.startswith('_'):
|
|
filename = filename[1:]
|
|
answer_dirs.append(filename+'/')
|
|
else:
|
|
answer.append(filename)
|
|
return answer, answer_dirs
|
|
CHEAT_SHEETS_TOPICS, CHEAT_SHEETS_DIRS = update_cheat_sheets_topics()
|
|
|
|
ANSI_ESCAPE = re.compile(r'(\x9B|\x1B\[)[0-?]*[ -\/]*[@-~]')
|
|
def remove_ansi(sometext):
|
|
return ANSI_ESCAPE.sub('', sometext)
|
|
|
|
def html_wrapper(data):
|
|
p = Popen([ "bash", ANSI2HTML, "--palette=solarized", "--bg=dark" ], stdin=PIPE, stdout=PIPE, stderr=PIPE)
|
|
data = data.encode('utf-8')
|
|
stdout, stderr = p.communicate(data)
|
|
if p.returncode != 0:
|
|
error(stdout + stderr)
|
|
return stdout.decode('utf-8')
|
|
|
|
#
|
|
#
|
|
#
|
|
|
|
cached_topics_list = [[]]
|
|
def get_topics_list(skip_dirs=False, skip_internal=False):
|
|
"""
|
|
List of topics returned on /:list
|
|
"""
|
|
|
|
if cached_topics_list[0] != []:
|
|
return cached_topics_list[0]
|
|
|
|
answer = CHEAT_TOPICS + TLDR_TOPICS + CHEAT_SHEETS_TOPICS
|
|
answer = sorted(set(answer))
|
|
|
|
# doing it in this strange way to save the order of the topics
|
|
for topic in get_learnxiny_list():
|
|
if topic not in answer:
|
|
answer.append(topic)
|
|
|
|
if not skip_dirs:
|
|
answer += CHEAT_SHEETS_DIRS
|
|
if not skip_internal:
|
|
answer += INTERNAL_TOPICS
|
|
|
|
cached_topics_list[0] = answer
|
|
return answer
|
|
|
|
def get_topics_dirs():
|
|
return set([x.split('/', 1)[0] for x in get_topics_list() if '/' in x])
|
|
|
|
#
|
|
#
|
|
#
|
|
|
|
def get_topic_type(topic):
|
|
if topic == "":
|
|
return "search"
|
|
|
|
if topic.startswith(":"):
|
|
return "internal"
|
|
if '/' in topic:
|
|
topic_type, topic_name = topic.split('/', 1)
|
|
if topic_type in get_topics_dirs() and topic_name in [':list']:
|
|
return "internal"
|
|
if is_valid_learnxy(topic):
|
|
return 'learnxiny'
|
|
if topic in CHEAT_SHEETS_TOPICS:
|
|
return "cheat.sheets"
|
|
if topic.rstrip('/') in CHEAT_SHEETS_DIRS and topic.endswith('/'):
|
|
return "cheat.sheets dir"
|
|
if topic in CHEAT_TOPICS:
|
|
return "cheat"
|
|
if topic in TLDR_TOPICS:
|
|
return "tldr"
|
|
return "unknown"
|
|
|
|
#
|
|
# Various cheat sheets getters
|
|
#
|
|
|
|
def get_internal(topic):
|
|
if '/' in topic:
|
|
topic_type, topic_name = topic.split('/', 1)
|
|
if topic_name == ":list":
|
|
topic_list = [x[len(topic_type)+1:]
|
|
for x in get_topics_list()
|
|
if x.startswith(topic_type + "/")]
|
|
return "\n".join(topic_list)+"\n"
|
|
|
|
if topic == ":list":
|
|
return "\n".join(x for x in get_topics_list())
|
|
|
|
if topic == ':styles':
|
|
return "\n".join(COLOR_STYLES)
|
|
|
|
if topic in INTERNAL_TOPICS:
|
|
return open(os.path.join(MYDIR, "share", topic[1:]+".txt"), "r").read()
|
|
|
|
return ""
|
|
|
|
def get_tldr(topic):
|
|
cmd = ["tldr", topic]
|
|
p = Popen(cmd, stdout=PIPE, stderr=PIPE)
|
|
answer = p.communicate()[0]
|
|
|
|
fixed_answer = []
|
|
for line in answer.splitlines():
|
|
line = line[2:]
|
|
if line.startswith('-'):
|
|
line = '# '+line[2:]
|
|
elif line == "":
|
|
pass
|
|
elif not line.startswith(' '):
|
|
line = "# "+line
|
|
|
|
fixed_answer.append(line)
|
|
|
|
answer = "\n".join(fixed_answer) + "\n"
|
|
return answer.decode('utf-8')
|
|
|
|
def get_cheat(topic):
|
|
cmd = ["cheat", topic]
|
|
p = Popen(cmd, stdout=PIPE, stderr=PIPE)
|
|
answer = p.communicate()[0].decode('utf-8')
|
|
return answer
|
|
|
|
def get_cheat_sheets(topic):
|
|
"""
|
|
Get the cheat sheet topic from the own repository (cheat.sheets).
|
|
It's possible that topic directory starts with omited underscore
|
|
"""
|
|
filename = PATH_CHEAT_SHEETS + "%s" % topic
|
|
if not os.path.exists(filename):
|
|
filename = PATH_CHEAT_SHEETS + "_%s" % topic
|
|
return open(filename, "r").read().decode('utf-8')
|
|
|
|
def get_cheat_sheets_dir(topic):
|
|
answer = []
|
|
for f_name in glob.glob(PATH_CHEAT_SHEETS + "%s/*" % topic.rstrip('/')):
|
|
answer.append(os.path.basename(f_name))
|
|
topics = sorted(answer)
|
|
return "\n".join(topics) + "\n"
|
|
|
|
def get_unknown(topic):
|
|
topics_list = get_topics_list()
|
|
if topic.startswith(':'):
|
|
topics_list = [x for x in topics_list if x.startswith(':')]
|
|
else:
|
|
topics_list = [x for x in topics_list if not x.startswith(':')]
|
|
|
|
possible_topics = process.extract(topic, topics_list, scorer=fuzz.ratio)[:3]
|
|
possible_topics_text = "\n".join([(" * %s %s" % x) for x in possible_topics])
|
|
return """
|
|
Unknown topic.
|
|
Do you mean one of these topics may be?
|
|
|
|
%s
|
|
""" % possible_topics_text
|
|
|
|
TOPIC_GETTERS = (
|
|
("cheat.sheets" , get_cheat_sheets),
|
|
("cheat.sheets dir" , get_cheat_sheets_dir),
|
|
("tldr" , get_tldr),
|
|
("internal" , get_internal),
|
|
("cheat" , get_cheat),
|
|
("learnxiny" , get_learnxiny),
|
|
("unknown" , get_unknown),
|
|
)
|
|
|
|
#
|
|
#
|
|
#
|
|
|
|
def split_paragraphs(text):
|
|
answer = []
|
|
paragraph = ""
|
|
for line in text.splitlines():
|
|
if line == "":
|
|
answer.append(paragraph)
|
|
paragraph = ""
|
|
else:
|
|
paragraph += line+"\n"
|
|
answer.append(paragraph)
|
|
|
|
return answer
|
|
|
|
def paragraph_contains(paragraph, keyword, insensitive=False, word_boundaries=True):
|
|
"""
|
|
Several keywords can be joined together using ~
|
|
For example: ~ssh~passphrase
|
|
"""
|
|
answer = True
|
|
|
|
if '~' in keyword:
|
|
keywords = keyword.split('~')
|
|
else:
|
|
keywords = [keyword]
|
|
|
|
for keyword in keywords:
|
|
regex = re.escape(keyword)
|
|
if not word_boundaries:
|
|
regex = r"\b%s\b" % keyword
|
|
|
|
if insensitive:
|
|
answer = answer and bool(re.search(regex, paragraph, re.IGNORECASE))
|
|
else:
|
|
answer = answer and bool(re.search(regex, paragraph))
|
|
|
|
return answer
|
|
|
|
def join_paragraphs(paragraphs):
|
|
answer = "\n".join(paragraphs)
|
|
return answer
|
|
|
|
def get_answer(topic, keyword, options=""):
|
|
"""
|
|
Find cheat sheet for the topic.
|
|
If not keyword, return answer.
|
|
Otherwise cut the paragraphs cotaining keywords.
|
|
|
|
Args:
|
|
topic (str): the name of the topic of the cheat sheet
|
|
keyword (str): the name of the keywords to search in the cheat sheets
|
|
|
|
Returns:
|
|
string: the cheat sheet
|
|
"""
|
|
|
|
answer = None
|
|
|
|
# checking if the answer is in the cache
|
|
if topic != "":
|
|
answer = REDIS.get(topic)
|
|
if answer:
|
|
answer = answer.decode('utf-8')
|
|
|
|
# if answer was not found in the cache
|
|
# try to find it in one of the repositories
|
|
if not answer:
|
|
topic_type = get_topic_type(topic)
|
|
for topic_getter_type, topic_getter in TOPIC_GETTERS:
|
|
if topic_type == topic_getter_type:
|
|
answer = topic_getter(topic)
|
|
break
|
|
if not answer:
|
|
answer = get_unknown(topic)
|
|
|
|
# saving answers in the cache
|
|
if topic_type not in ["search", "internal", "unknown"]:
|
|
REDIS.set(topic, answer)
|
|
|
|
if not keyword:
|
|
return answer
|
|
|
|
insensitive = 'i' in options
|
|
word_boundaries = 'b' in options
|
|
|
|
paragraphs = split_paragraphs(answer)
|
|
paragraphs = [ p for p in paragraphs
|
|
if paragraph_contains(
|
|
p, keyword,
|
|
insensitive=insensitive,
|
|
word_boundaries=word_boundaries) ]
|
|
if len(paragraphs) == 0:
|
|
return ""
|
|
|
|
answer = join_paragraphs(paragraphs)
|
|
return answer
|
|
|
|
def find_answer_by_keyword(directory, keyword, options=""):
|
|
recursive = 'r' in options
|
|
|
|
answer_paragraphs = []
|
|
for topic in get_topics_list(skip_internal=True, skip_dirs=True):
|
|
# skip these pages, don't show them in search
|
|
if topic in [':firstpage']:
|
|
continue
|
|
|
|
if not topic.startswith(directory):
|
|
continue
|
|
|
|
subtopic = topic[len(directory):]
|
|
if not recursive and '/' in subtopic:
|
|
continue
|
|
|
|
answer = get_answer(topic, keyword, options=options)
|
|
if answer:
|
|
answer_paragraphs.append((topic, answer))
|
|
|
|
MAX_SEARCH_LEN = 20
|
|
if len(answer_paragraphs) > MAX_SEARCH_LEN:
|
|
answer_paragraphs.append(("LIMITED", "LIMITED TO %s ANSWERS" % MAX_SEARCH_LEN))
|
|
break
|
|
|
|
return answer_paragraphs
|
|
|
|
#
|
|
|
|
def save_cheatsheet(topic_name, cheatsheet):
|
|
|
|
nonce = ''.join(random.choice(string.ascii_uppercase + string.digits) for _ in range(9))
|
|
filename = topic_name.replace('/', '.') + "." + nonce
|
|
filename = os.path.join(PATH_CHEAT_SHEETS_SPOOL, filename)
|
|
|
|
open(filename, 'w').write(cheatsheet)
|
|
#
|
|
#
|
|
#
|
|
|
|
def colorize_internal(topic, answer, html_needed):
|
|
|
|
def colorize_line(line):
|
|
if line.startswith('T'):
|
|
line = colored.fg("grey_62") + line + colored.attr('reset')
|
|
line = re.sub(r"\{(.*?)\}", colored.fg("orange_3") + r"\1"+colored.fg('grey_35'), line)
|
|
return line
|
|
|
|
line = re.sub(r"\[(F.*?)\]", colored.bg("black") + colored.fg("cyan") + r"[\1]"+colored.attr('reset'), line)
|
|
line = re.sub(r"\[(g.*?)\]", colored.bg("dark_gray") + colored.fg("grey_0") + r"[\1]"+colored.attr('reset'), line)
|
|
line = re.sub(r"\{(.*?)\}", colored.fg("orange_3") + r"\1"+colored.attr('reset'), line)
|
|
line = re.sub(r"<(.*?)>", colored.fg("cyan") + r"\1"+colored.attr('reset'), line)
|
|
return line
|
|
|
|
if topic in [':list', ':bash_completion']:
|
|
return answer
|
|
|
|
if topic == ':firstpage':
|
|
lines = answer.splitlines()
|
|
answer_lines = lines[:9]
|
|
answer_lines.append(colored.fg('grey_35')+lines[9]+colored.attr('reset'))
|
|
for line in lines[10:]:
|
|
answer_lines.append(colorize_line(line))
|
|
if html_needed:
|
|
answer_lines = answer_lines[:-2]
|
|
answer = "\n".join(answer_lines) + "\n"
|
|
|
|
return answer
|
|
|
|
def github_button(topic_type):
|
|
repository = {
|
|
"cheat.sheets" : 'chubin/cheat.sheets',
|
|
"cheat.sheets dir" : 'chubin/cheat.sheets',
|
|
"tldr" : 'tldr-pages/tldr',
|
|
"cheat" : 'chrisallenlane/cheat',
|
|
"learnxiny" : 'adambard/learnxinyminutes-docs',
|
|
"internal" : '',
|
|
"search" : '',
|
|
"unknown" : '',
|
|
}
|
|
|
|
full_name = repository.get(topic_type, '')
|
|
if not full_name:
|
|
return ''
|
|
|
|
short_name = full_name.split('/',1)[1]
|
|
button = (
|
|
"<!-- Place this tag where you want the button to render. -->"
|
|
'<a aria-label="Star %(full_name)s on GitHub" data-count-aria-label="# stargazers on GitHub"'
|
|
' data-count-api="/repos/%(full_name)s#stargazers_count"'
|
|
' data-count-href="/%(full_name)s/stargazers"'
|
|
' data-icon="octicon-star"'
|
|
' href="https://github.com/%(full_name)s"'
|
|
' class="github-button">%(short_name)s</a>'
|
|
) % locals()
|
|
return button
|
|
|
|
#
|
|
|
|
def rewrite_aliases(word):
|
|
if word == ':bash.completion':
|
|
return ':bash_completion'
|
|
return word
|
|
|
|
def cheat_wrapper(query, request_options=None, html=False):
|
|
|
|
#
|
|
# at the moment, we just remove trailing slashes
|
|
# so queries python/ and python are equal
|
|
#
|
|
query = query.rstrip('/')
|
|
|
|
query = rewrite_aliases(query)
|
|
|
|
highlight = not bool(request_options and request_options.get('no-terminal'))
|
|
color_style = request_options.get('style', '')
|
|
if color_style not in COLOR_STYLES:
|
|
color_style = ''
|
|
|
|
keyword = None
|
|
if '~' in query:
|
|
topic = query
|
|
pos = topic.index('~')
|
|
keyword = topic[pos+1:]
|
|
topic = topic[:pos]
|
|
|
|
options = ""
|
|
if '/' in keyword:
|
|
options = keyword[::-1]
|
|
options = options[:options.index('/')]
|
|
keyword = keyword[:-len(options)-1]
|
|
|
|
answers = find_answer_by_keyword(topic, keyword, options=options)
|
|
search_mode = True
|
|
else:
|
|
answers = [(query, get_answer(query, keyword))]
|
|
search_mode = False
|
|
|
|
|
|
found = True # if the page was found in the database
|
|
editable = False # can generated page be edited on github (only cheat.sheets pages can)
|
|
result = ""
|
|
for topic, answer in answers:
|
|
|
|
if topic == 'LIMITED':
|
|
result += colored.bg('dark_goldenrod') + colored.fg('yellow_1') + ' ' + answer + ' ' + colored.attr('reset') + "\n"
|
|
break
|
|
|
|
if topic in [":list", ":bash_completion"]:
|
|
highlight = False
|
|
|
|
topic_type = get_topic_type(topic)
|
|
if topic_type == 'unknown':
|
|
found = False
|
|
|
|
if highlight:
|
|
#if topic_type.endswith(" dir"):
|
|
# pass
|
|
|
|
if topic_type == "internal":
|
|
answer = colorize_internal(topic, answer, html)
|
|
else:
|
|
color_style = color_style or "native"
|
|
lexer = BashLexer
|
|
for lexer_name, lexer_value in LEXER.items():
|
|
if topic.startswith("%s/" % lexer_name):
|
|
color_style = color_style or "monokai"
|
|
if lexer_name == 'php':
|
|
answer = "<?\n%s?>\n" % answer
|
|
lexer = lexer_value
|
|
break
|
|
|
|
formatter = Terminal256Formatter(style=color_style)
|
|
answer = pygments_highlight(answer, lexer(), formatter)
|
|
|
|
if topic_type == "cheat.sheets":
|
|
editable = True
|
|
|
|
if search_mode:
|
|
if highlight:
|
|
result += "\n%s%s %s %s%s\n" % (colored.bg('dark_gray'), colored.attr("res_underlined"), topic, colored.attr("res_underlined"), colored.attr('reset'))
|
|
else:
|
|
result += "[%s]" % topic
|
|
|
|
result += answer
|
|
|
|
if search_mode:
|
|
result = result[1:]
|
|
editable = False
|
|
repository_button = ''
|
|
else:
|
|
repository_button = github_button(topic_type)
|
|
|
|
if html:
|
|
result = result + "\n$"
|
|
result = html_wrapper(result)
|
|
title = "<title>cheat.sh/%s</title>" % topic
|
|
# title += '\n<link rel="stylesheet" href="/files/awesomplete.css" />script src="/files/awesomplete.min.js" async></script>'
|
|
# submit button: thanks to http://stackoverflow.com/questions/477691/
|
|
submit_button = '<input type="submit" style="position: absolute; left: -9999px; width: 1px; height: 1px;" tabindex="-1" />'
|
|
topic_list = ('<datalist id="topics">%s</datalist>'
|
|
% ("\n".join("<option value='%s'></option>" % x for x in get_topics_list())))
|
|
|
|
curl_line = "<span class='pre'>$ curl cheat.sh/</span>"
|
|
if query == ':firstpage':
|
|
query = ""
|
|
form_html = '<form action="/" method="GET"/>%s%s<input type="text" value="%s" name="topic" list="topics" autofocus autocomplete="off"/>%s</form>' % (submit_button, curl_line, query, topic_list)
|
|
|
|
edit_button = ''
|
|
if editable:
|
|
edit_page_link = 'https://github.com/chubin/cheat.sheets/edit/master/sheets/' + topic
|
|
edit_button = '<pre style="position:absolute;padding-left:40em;overflow:visible;height:0;">[<a href="%s" style="color:cyan">edit</a>]</pre>' % edit_page_link
|
|
result = re.sub("<pre>", edit_button + form_html + "<pre>", result)
|
|
result = re.sub("<head>", "<head>" + title, result)
|
|
if not request_options.get('quiet'):
|
|
result = result.replace('</body>', TWITTER_BUTTON + GITHUB_BUTTON + repository_button + GITHUB_BUTTON_FOOTER + '</body>')
|
|
|
|
return result, found
|
|
|