def zip_dict(dst_fname):
dir_fpath, basename = os.path.split(dst_fname)
root_dir, dir_fname = os.path.split(dir_fpath)
# uzado de zip ne funkciigas, rompas Colordict - t.e. la lasta komencas blinki senfine
#fmt = "zip"
fmt = "gztar"
# se vortara datumo estas ĝuste en arkivo, ne en ia dosierujo, do CSS/figuroj estas ne trovataj
save_without_folder = False # True #
if save_without_folder:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname), fmt, dir_fpath)
else:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname), fmt, root_dir, base_dir=dir_fname)
ifo_fname = os.path.splitext(dst_fname)[0] + ".ifo"
with open(ifo_fname) as ifo_f:
properties = {}
for line in ifo_f:
lst = line.split("=")
if len(lst) >= 2:
key, value = lst[0].strip(), lst[1].strip()
if key and value:
properties[key] = value
words_cnt = int(properties.get("wordcount"))
synwordcount = properties.get("synwordcount")
if synwordcount:
words_cnt += int(synwordcount)
fname = os.path.basename(fname)
# du spacetoj fine estas por Markdown liniavanco
print("http://new.bombono.org/download/revo/%(fname)s\t%(words_cnt)s " % locals())
def install(rpkg_tar_gz, dir_fpath):
assert os.path.isdir(dir_fpath)
def set_current(to_last):
src = "last" if to_last else "previous"
# :TRICKY: для переопределения сущ. ссылки нужно одновременно -f и -n ,-
# артефакт *nix
call_cmd("ln -s -f -n %s current" % src, cwd=dir_fpath)
last_fpath = o_p.join(dir_fpath, "last")
curr_fpath = o_p.join(dir_fpath, "current")
if os.path.exists(last_fpath):
curr_exists = False
if os.path.exists(curr_fpath):
assert os.path.islink(curr_fpath)
curr_exists = True
prev_fpath = o_p.join(dir_fpath, "previous")
if os.path.exists(prev_fpath):
shutil.rmtree(prev_fpath)
shutil.move(last_fpath, prev_fpath)
if curr_exists:
set_current(False)
os.mkdir(last_fpath)
call_cmd("tar -xzf %(rpkg_tar_gz)s -C %(last_fpath)s" % locals())
set_current(True)
def install(rpkg_tar_gz, dir_fpath):
assert os.path.isdir(dir_fpath)
def set_current(to_last):
src = "last" if to_last else "previous"
# :TRICKY: для переопределения сущ. ссылки нужно одновременно -f и -n ,-
# артефакт *nix
call_cmd("ln -s -f -n %s current" % src, cwd=dir_fpath)
last_fpath = o_p.join(dir_fpath, "last")
curr_fpath = o_p.join(dir_fpath, "current")
if os.path.exists(last_fpath):
curr_exists = False
if os.path.exists(curr_fpath):
assert os.path.islink(curr_fpath)
curr_exists = True
prev_fpath = o_p.join(dir_fpath, "previous")
if os.path.exists(prev_fpath):
shutil.rmtree(prev_fpath)
shutil.move(last_fpath, prev_fpath)
if curr_exists:
set_current(False)
os.mkdir(last_fpath)
call_cmd("tar -xzf %(rpkg_tar_gz)s -C %(last_fpath)s" % locals())
set_current(True)
def append_national_article(lang, names, txt):
o_p_article, dst_fname = dictionaries.get(lang, (None, None))
if o_p_article is None:
dict_fpath = o_p.join(dct_prefix, "REVO_%(lang)s" % locals())
# :REFACTOR:
dst_fname = o_p.join(dict_fpath, "REVO-%(lang)s.txt" % locals())
o_p_article = create_dictionary(dst_fname, "revo-traduko.css")
dictionaries[lang] = o_p_article, dst_fname
copy_prj_fname("sample/revo/eo-nacia/revo-traduko.css", o_p.join(dict_fpath, "res/revo-traduko.css"))
o_p_article(names, txt)
def make_rel2abs(rel_fpath, stack_depth=1):
"""
Построить абсолютный путь относительно пути модуля
Значение stack_depth такое же, как и для get_mdir()
"""
dir_fpath = get_mdir(stack_depth+1)
return o_p.join(dir_fpath, rel_fpath)
def append_national_article(lang, names, txt):
o_p_article, dst_fname = dictionaries.get(lang, (None, None))
if o_p_article is None:
dict_fpath = o_p.join(dct_prefix,
"REVO_%(lang)s" % locals())
# :REFACTOR:
dst_fname = o_p.join(dict_fpath,
"REVO-%(lang)s.txt" % locals())
o_p_article = create_dictionary(dst_fname,
"revo-traduko.css")
dictionaries[lang] = o_p_article, dst_fname
copy_prj_fname(
"sample/revo/eo-nacia/revo-traduko.css",
o_p.join(dict_fpath, "res/revo-traduko.css"))
o_p_article(names, txt)
def request_chunk(rtp_db, startstamp):
payload = b''
# список чанков ровно с таким UTC (на единицу больше уже не подходит)
names = request_names(rtp_db, startstamp, 1)
if names:
with open(o_p.join(rtp_db2dir(rtp_db), names[0]), "rb") as f:
payload = f.read()
return payload
def zip_dict(dst_fname):
dir_fpath, basename = os.path.split(dst_fname)
root_dir, dir_fname = os.path.split(dir_fpath)
# uzado de zip ne funkciigas, rompas Colordict - t.e. la lasta komencas blinki senfine
#fmt = "zip"
fmt = "gztar"
# se vortara datumo estas ĝuste en arkivo, ne en ia dosierujo, do CSS/figuroj estas ne trovataj
save_without_folder = False # True #
if save_without_folder:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname),
fmt, dir_fpath)
else:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname),
fmt,
root_dir,
base_dir=dir_fname)
ifo_fname = os.path.splitext(dst_fname)[0] + ".ifo"
with open(ifo_fname) as ifo_f:
properties = {}
for line in ifo_f:
lst = line.split("=")
if len(lst) >= 2:
key, value = lst[0].strip(), lst[1].strip()
if key and value:
properties[key] = value
words_cnt = int(properties.get("wordcount"))
synwordcount = properties.get("synwordcount")
if synwordcount:
words_cnt += int(synwordcount)
fname = os.path.basename(fname)
# du spacetoj fine estas por Markdown liniavanco
print(
"http://new.bombono.org/download/revo/%(fname)s\t%(words_cnt)s " %
locals())
def create_vip_vortaro(dst_fname, remove_srcfile=True):
use_external_css = False
css_txt = read_all_file(o_p.join(project_dir_fname, "sample/vip/vip.css"))
def parse_dictionary(on_parsed_article):
src_fdir = make_dpath("articles")
# :REFACTOR:
for fname in os.listdir(src_fdir):
txt = read_all_file(os.path.join(src_fdir, fname))
all_names = []
vip_href_prefix = "?w="
fixed_text = []
php_len = len(vip_href_prefix)
def on_article(article):
all_names.extend(get_search_words(article))
for a in iter_tags(article, "a"):
href = a.get("href")
assert href.startswith(vip_href_prefix)
# подсказка по bword:
# http://siripong-english.blogspot.ru/2011/04/custom-dictionary-w-stardict-in-babylon.html
a.set("href", "bword://" + href[php_len:])
fixed_text.append(gen_html_text(article))
process_vip_text(txt, on_article)
txt = fixed_text[0]
css_txt
article_txt = txt # "<style>%(css_txt)s</style>%(txt)s" % locals()
on_parsed_article(all_names, article_txt)
make_kondratjev.make_dictionary_ex(parse_dictionary, dst_fname, css_text=None if use_external_css else css_txt,
remove_srcfile=remove_srcfile, is_html=True)
def make_path(name):
return o_p.join(os.path.dirname(__file__), name)
pat = re.compile(r"^#EXT-X-STREAM-INF.+BANDWIDTH=(?P<bandwidth>\d+).*(?:\n|\r\n?)(?P<stream>.+)", re.MULTILINE)
dst_fname = {
"m3u8": "vlc.m3u8",
"html": "tv_bl.html",
"xspf": "vlc.xspf"
}[fmt]
req_clns = ["ts_port"]
for cnxt in rewrite_channels(dst_fname, req_clns, fmt=fmt):
# :TRICKY: своей колонки нет
hls_idx = cnxt.clns["ts_port"] + 1
url = cnxt.row[hls_idx]
if url.startswith("http://"):
print(name, url)
try:
with contextlib.closing(get_url.get_url(url)) as pf:
txt = pf.read()
except get_url.URLError:
pass
else:
max_bw, max_url = 0, None
for m in pat.finditer(txt):
bw = int(m.group('bandwidth'))
if not max_bw or max_bw < bw:
max_bw = bw
max_url = m.group('stream')
assert max_url
max_url = o_p.join(os.path.dirname(url), max_url)
write_channel(cnxt, max_url)
re.MULTILINE)
dst_fname = {
"m3u8": "vlc.m3u8",
"html": "tv_bl.html",
"xspf": "vlc.xspf"
}[fmt]
req_clns = ["ts_port"]
for cnxt in rewrite_channels(dst_fname, req_clns, fmt=fmt):
# :TRICKY: своей колонки нет
hls_idx = cnxt.clns["ts_port"] + 1
url = cnxt.row[hls_idx]
if url.startswith("http://"):
print(name, url)
try:
with contextlib.closing(get_url.get_url(url)) as pf:
txt = pf.read()
except get_url.URLError:
pass
else:
max_bw, max_url = 0, None
for m in pat.finditer(txt):
bw = int(m.group('bandwidth'))
if not max_bw or max_bw < bw:
max_bw = bw
max_url = m.group('stream')
assert max_url
max_url = o_p.join(os.path.dirname(url), max_url)
write_channel(cnxt, max_url)
def make_path(name):
return o_p.join(os.path.dirname(__file__), name)
def open_xml_article(xml_fname):
xml_fname = o_p.join(xml_fpath, xml_fname)
tree = open_xml_tree(xml_fname)
return tree
def main():
dct_prefix = "/home/ilya/.stardict/dic/esperanto"
dst_fname = o_p.join(dct_prefix, "REVO_Eksplika/Eksplika-REVO.txt")
# :REFACTOR:
dirname = os.path.dirname
# :REFACTOR: realpath() - apliki tion al ĉiuj uzoj
prj_fdir = dirname(os.path.realpath(__file__))
import shutil
def copy_prj_fname(prj_fname, dst_fpath):
o_p.force_makedirs(os.path.dirname(dst_fpath))
shutil.copy(o_p.join(prj_fdir, prj_fname), dst_fpath)
prefix_eoru = dirname(dirname(prj_fdir))
unpacked_revo = o_p.join(prefix_eoru, "stuff/revo/revo")
dictionaries = {}
with make_gen_accumulator() as add_gen:
def create_dictionary(dst_fname, css_link=None):
remove_srcfile = True # False #
on_article = add_gen(
make_kondratjev.dictionary_generator(
dst_fname,
css_text=None,
is_html=True,
remove_srcfile=remove_srcfile))
if css_link:
orig_on_article = on_article
def on_article(key_names, txt):
css_link
txt = """<link href="%(css_link)s" rel="stylesheet" type="text/css" />%(txt)s""" % locals(
)
return orig_on_article(key_names, txt)
return on_article
on_explika_article = create_dictionary(dst_fname, "revo.css")
res_fdir = o_p.join(dirname(dst_fname), "res")
copy_prj_fname("sample/revo/revo.css", o_p.join(res_fdir, "revo.css"))
# kopias figurojn por beleco
dst_smb = o_p.join(res_fdir, "smb")
if not o_p.exists(dst_smb):
shutil.copytree(o_p.join(unpacked_revo, "smb"), dst_smb)
xml_fpath = o_p.join(unpacked_revo, "xml")
def open_xml_article(xml_fname):
xml_fname = o_p.join(xml_fpath, xml_fname)
tree = open_xml_tree(xml_fname)
return tree
def fname2prefix(src_fname):
return o_p.without_ext(src_fname)
prefix_dct = {}
def get_words(prefix):
words = prefix_dct.get(prefix)
if words is None:
words = prefix_dct[prefix] = []
tree = open_xml_article(prefix + ".xml")
for drv, headwords in for_drv_words_headwords(tree):
words.extend(calc_words(headwords))
#print(words)
#print(rvut_definitions.get_translations(drv).get("de"))
#print()
return words
fname_lst = os.listdir(xml_fpath)
if False: # True: #
fname_lst = [
"ten.xml",
"distin.xml",
"apenau.xml", # <trd> in <subdrv>
"pri.xml", # artikolo sen <drv>
"sur.xml", # <ekz> ne en <(sub)snc>, sed en <subart>
"al.xml", # <trdgrp> ĝuste en <art>
"stift.xml", # kaj <ekz> ankaŭ en <art>
"lima.xml", # перевод относился к <kap>, хотя был внутри текста (гад с 'la') - и таких статей много
"kverk.xml", # diversaj homaj eraroj
"jxak1.xml",
"anim.xml", # <ekz> sen <ind>
"blank.xml", #
"milv.xml", # <bld> anstataŭ <>ekz
"hel.xml", # trdgrp en <dif>
"hazard.xml", # malplena trd etikedo
"iks.xml", # vortoj kun signo '|'
]
for src_fname in fname_lst:
prefix = fname2prefix(src_fname)
all_names = get_words(prefix)
html_fname = o_p.join(unpacked_revo, "art", prefix + ".html")
body = make_wells.get_html_body(html_fname, False)
h1 = body.find("h1")
hr = body.find("hr")
div = etree.Element("div")
el = h1.getnext()
while el != hr:
div.append(el)
el = h1.getnext()
def append_sub(name):
sub_el = body.find("div[@class='%(name)s']" % locals())
if not (sub_el is None):
div.append(etree.Element("hr"))
div.append(sub_el)
append_sub("fontoj")
append_sub("notoj")
# renovigas referencojn en stilo
# kapt.html#kapt.0i => bword://kapti#kapt.0i
for lnk in parse_vip.iter_tags(div, "a"):
href = lnk.get("href")
if href:
m = re.match(r"(?P<lnk_fname>[^/]+\.html)#(?P<anchor>.+)$",
href)
if m:
lnk_fname, anchor = m.group("lnk_fname"), m.group(
"anchor")
lnk_word = get_words(fname2prefix(lnk_fname))[0]
# GD ne atentas #anchor, ColorDict - eĉ rifuzas sekvi la ligilon
#lnk.set("href", "bword://%(lnk_word)s#%(anchor)s" % locals())
lnk.set("href",
"bword://%(lnk_word)s#%(anchor)s" % locals())
# :REFACTOR:
for img in parse_vip.iter_tags(div, "img"):
src = img.get("src")
if src:
# egala funkciado por Goldendict (GD) k ColorDict (CD)
m = re.match(r"^\.\./", src)
if m:
img.set("src", src[3:])
txt = parse_vip.gen_html_text(div)
#print(txt)
on_explika_article(all_names, txt)
# eo-nacia vortaro
national_dct = {}
tree = open_xml_article(src_fname)
def append_translations(translations, src_trs):
for lang, lst in src_trs.items():
translations[lang] = lst + translations.setdefault(
lang, [])
used_tr_nodes = {}
national_headwords = {}
def get_count_translations(node):
res = rvut_definitions.get_translations(node)
# hazard.xml havas malplena tradukojn
clean_res = {}
for lang, lst in res.items():
lst = list(filter(bool, lst))
if lst:
clean_res[lang] = lst
res = clean_res
append_translations(national_headwords, res)
# :REFACTOR:
for trd in node.findall('trd'):
used_tr_nodes[trd] = True
for trdp in node.findall('trdgrp'):
used_tr_nodes[trdp] = True
for trd in trdp.findall('trd'):
used_tr_nodes[trd] = True
return res
def iterate_translations(translations, sub_node_dct,
numerator_func, final_sep):
for lang in sub_node_dct.keys() | translations.keys():
yield lang, gen_trans_text(lang, sub_node_dct,
numerator_func, translations,
final_sep)
def notify_node(warning_txt, node):
print(warning_txt, src_fname, parse_vip.gen_html_text(node))
# :TRICKY: plej simpla maniero por kalkuki jam traktitajn nodojn
ekz_node_set = set()
def find_ekz_translations(ekz_dct, node, flat_translations):
#for trd in parse_vip.iter_tags(node, "ekz/trd|trdgrp"):
def trd_iter(ekz_name, name):
return parse_vip.iter_tags(
node, "%(ekz_name)s/%(name)s" % locals())
def trd_iters(ekz_name):
return trd_iter(ekz_name,
"trd"), trd_iter(ekz_name, "trdgrp")
for trd in itertools.chain(*(trd_iters("ekz") +
trd_iters("bld"))):
ekz = trd.getparent()
if ekz in ekz_node_set:
continue
else:
ekz_node_set.add(ekz)
def make_orig_txt(ind_node):
return ', '.join(
rvut_words.get_words_from_kap(ind_node))
ind_node = ekz.find('ind')
if ind_node is None:
# kalkulas orig_txt mem, kolektante ĉiujn etikedojn ĝis apero de trd aŭ trdgrp
# anim.xml:
# <ekz>
# <tld/>ita parolado<fnt>K</fnt>,
# <trd lng="hu">lelkes beszéd</trd>
# </ekz>
ind_node = etree.Element("ind")
ind_node.text = ekz.text
for child in ekz.getchildren():
if child.tag in ["trd", "trdgrp"]:
break
else:
child = copy.deepcopy(child)
ind_node.append(child)
tree.append(ind_node)
orig_txt = make_orig_txt(ind_node)
ind_node.getparent().remove(ind_node)
else:
orig_txt = make_orig_txt(ind_node)
for lang, tr_lst in get_count_translations(ekz).items():
# :REFACTOR:
lst = ekz_dct.setdefault(lang, [])
tr_lst = ", ".join(tr_lst)
ekz_txt = "<i><b>%(orig_txt)s</b>: %(tr_lst)s</i>" % locals(
)
lst.append(ekz_txt)
#return
# :TRICKY: iuj <trd> kumulas tradukon mem k indikon de originala nomo (Latina prezipe) =>
# nur <trd> povas esti tia, ne <trdgrp>, ĉar tio estas perokula etikedo (angla - tag)
# malĝuste - hel.xml!
rest_translations = {}
for trd in parse_vip.iter_tags(node, "trd"):
if trd not in used_tr_nodes:
par_node = trd.getparent()
if par_node.tag == "trdgrp":
lang = par_node.get("lng")
used_tr_nodes[par_node] = True
else:
lang = trd.get("lng")
foreign_word = rvut_flatten.flatten_node(trd)
if foreign_word:
# :REFACTOR:
rest_translations.setdefault(
lang, []).append(foreign_word)
# :REFACTOR:
used_tr_nodes[trd] = True
append_translations(flat_translations, rest_translations)
append_translations(national_headwords, rest_translations)
def append_ekz_translations(dct, ekz_dct):
# :TRICKY: ke lasi subsnc_dct simplan kaj ne ŝanĝi iterater_translations,
# do ĵuste aldonas ekzemplojn al lasta ero de subsnc_dct
for lang, ekz_lst in ekz_dct.items():
ekz_txt = "; ".join(ekz_lst)
lst = dct.setdefault(lang, [])
if lst:
lst[-1] += "; " + ekz_txt
else:
lst.append(ekz_txt)
def append_national_article(lang, names, txt):
o_p_article, dst_fname = dictionaries.get(lang, (None, None))
if o_p_article is None:
dict_fpath = o_p.join(dct_prefix,
"REVO_%(lang)s" % locals())
# :REFACTOR:
dst_fname = o_p.join(dict_fpath,
"REVO-%(lang)s.txt" % locals())
o_p_article = create_dictionary(dst_fname,
"revo-traduko.css")
dictionaries[lang] = o_p_article, dst_fname
copy_prj_fname(
"sample/revo/eo-nacia/revo-traduko.css",
o_p.join(dict_fpath, "res/revo-traduko.css"))
o_p_article(names, txt)
def append_row(translations, snc_dct, headwords, drv):
# sur.xml: <ekz> povas esti ekster ajna <snc>
ekz_dct = {}
find_ekz_translations(ekz_dct, drv, translations)
append_ekz_translations(translations, ekz_dct)
assert headwords
hw_txt = "<b>%s</b>" % "</b>, <b>".join(headwords)
typ = None
vspec = drv.find("gra/vspec")
if vspec is not None:
typ = vspec.text
if typ:
hw_txt = "%(hw_txt)s <i>%(typ)s</i>" % locals()
for lang, tr_txt in iterate_translations(
translations, snc_dct, arab_num, " <b>|</b> "):
opa_args = national_dct.setdefault(lang, ([], []))
names, txt = opa_args
names.extend(calc_words(headwords))
row_txt = """<div class="paragrafo">%(hw_txt)s %(tr_txt)s</div>""" % locals(
)
txt.append(row_txt)
# nacia-eo article
n_keywords = national_headwords.get(lang)
assert n_keywords
# devas purigi poste originalan n_keywords, ne ŝanĝitan
#n_keywords = [word.replace("|", "/") for word in n_keywords]
clean_keywords = [
word.replace("|", "/") for word in n_keywords
]
append_national_article(lang, clean_keywords, row_txt)
n_keywords.clear()
for drv, headwords in for_drv_words_headwords(tree):
#print(src_fname)
#print(translations)
#print()
def latin_num(i):
return "%(chr(ord('a') + i))s)" % s_.EvalFormat()
snc_dct = {}
ekz_snc_dct = {}
for snc in parse_vip.iter_tags(drv, "snc"):
subsnc_dct = {}
ekz_subsnc_dct = {}
for subsnc in parse_vip.iter_tags(snc, "subsnc"):
subsnc_translations = get_count_translations(subsnc)
for lang, tr_lst in subsnc_translations.items():
lst = subsnc_dct.setdefault(lang, [])
lst.append(", ".join(tr_lst))
find_ekz_translations(ekz_subsnc_dct, subsnc,
subsnc_dct)
append_ekz_translations(subsnc_dct, ekz_subsnc_dct)
for lang, tr_txt in iterate_translations(
get_count_translations(snc), subsnc_dct, latin_num,
"; "):
lst = snc_dct.setdefault(lang, [])
lst.append(tr_txt)
find_ekz_translations(ekz_snc_dct, snc, snc_dct)
def arab_num(i):
return "<b>%(i+1)s.</b>" % s_.EvalFormat()
append_ekz_translations(snc_dct, ekz_snc_dct)
def merge_trs(translations, drv):
src_trs = get_count_translations(drv)
append_translations(translations, src_trs)
# ankoraŭ estas iome da <subdrv> en <drv> => aldonu
translations = {}
for subdrv in parse_vip.iter_tags(drv, "subdrv"):
merge_trs(translations, subdrv)
# k subart ankaŭ eblas havi <trd> ĝuste en subart - sur.xml: hu => rá-
merge_trs(translations, drv)
append_row(translations, snc_dct, headwords, drv)
# :TRICKY: al.xml havas tradukojn ekstere subart, drv
art_node = tree.find("art")
append_row(get_count_translations(art_node), {},
find_kap_words(art_node), art_node)
for lang, opa_args in national_dct.items():
names, txt = opa_args
append_national_article(lang, names, "".join(txt))
strict_check = False # True #
def alarm_not_processed(trd):
is_ok = trd in used_tr_nodes
if not is_ok:
if strict_check:
assert is_ok
else:
notify_node("Not processed trd:", trd.getparent())
# kontrolo, ke ĉiuj nodoj estas traktita
for trd in parse_vip.iter_tags(tree, "trd"):
alarm_not_processed(trd)
for trd in parse_vip.iter_tags(tree, "trdgrp"):
alarm_not_processed(trd)
# zip'u vortarojn
revo_dicts_fpath = o_p.join(dirname(unpacked_revo), "revo-dicts")
o_p.force_makedirs(revo_dicts_fpath)
# shutil povas zipfile!
#import zipfile
print("\nAtingeblaj REVO vortaroj:")
def zip_dict(dst_fname):
dir_fpath, basename = os.path.split(dst_fname)
root_dir, dir_fname = os.path.split(dir_fpath)
# uzado de zip ne funkciigas, rompas Colordict - t.e. la lasta komencas blinki senfine
#fmt = "zip"
fmt = "gztar"
# se vortara datumo estas ĝuste en arkivo, ne en ia dosierujo, do CSS/figuroj estas ne trovataj
save_without_folder = False # True #
if save_without_folder:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname),
fmt, dir_fpath)
else:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname),
fmt,
root_dir,
base_dir=dir_fname)
ifo_fname = os.path.splitext(dst_fname)[0] + ".ifo"
with open(ifo_fname) as ifo_f:
properties = {}
for line in ifo_f:
lst = line.split("=")
if len(lst) >= 2:
key, value = lst[0].strip(), lst[1].strip()
if key and value:
properties[key] = value
words_cnt = int(properties.get("wordcount"))
synwordcount = properties.get("synwordcount")
if synwordcount:
words_cnt += int(synwordcount)
fname = os.path.basename(fname)
# du spacetoj fine estas por Markdown liniavanco
print(
"http://new.bombono.org/download/revo/%(fname)s\t%(words_cnt)s " %
locals())
zip_dict(dst_fname)
for lang, (func, dst_fname) in dictionaries.items():
zip_dict(dst_fname)
def main():
prefix = "/home/ilya/.stardict/dic/esperanto"
make_wells(15, 42, o_p.join(prefix, "wells-eo-en/wells-eo-en.txt"))
make_wells(44, 69, o_p.join(prefix, "wells-en-eo/wells-en-eo.txt"))
type=bool, default=False,
)
parser.add_argument("src_path", help="where is ffmpeg/libav source directory")
args = parser.parse_args()
is_libav = args.is_libav
src_dir = os.path.abspath(args.src_path)
if is_libav:
add_opts = ""
else:
if args.is_debug:
add_opts = "--disable-stripping"
else:
add_opts = ""
bld_dir = o_p.join(src_dir, "objs")
if os.path.exists(bld_dir):
import shutil
shutil.rmtree(bld_dir)
o_p.force_makedirs(bld_dir)
inst_dir = o_p.join(bld_dir, "inst")
enable_x264 = True
x264_support = "--enable-libx264 --enable-gpl" if enable_x264 else ""
if x264_support:
x264_support = " " + x264_support
debug_opts = "--disable-optimizations --extra-cflags='-O0 -g' " if args.is_debug else ""
call_cmd("../configure --prefix=%(inst_dir)s %(debug_opts)s--extra-ldflags='-Wl,-rpath=%(inst_dir)s/lib' \
def parse_dictionary(on_parsed_article):
dirname = os.path.dirname
prefix_eoru = dirname(dirname(dirname(__file__)))
unpacked_epub = o_p.join(prefix_eoru, "stuff/Wells/decrypted/depacked")
for num in range(first_num, last_num + 1):
#num = 29 # K
src_fname = o_p.join(unpacked_epub, "OEBPS/%03d.html" % num)
body = get_html_body(src_fname, True)
found_empty_p = False
# альтернативный способ - .getchildren() + проверка .tag
for p in body.iterfind(xhtml_tag("p")):
txt = p.text_content().strip()
if found_empty_p and txt:
# очередная статья
#print(txt)
radix = None
lst = []
def on_word(word):
# <b>Kaboverd/o</b><b> </b>Cape Verde
if word:
lst.append(word)
key_elements = list(parse_vip.iter_tags(p, xhtml_tag("b")))
assert key_elements
for idx, el in enumerate(key_elements):
bold_txt = el.text_content().strip()
exceptions = [
"li diris, ~ ŝi atendas", # ke
"~e, ke", # kondiĉo
]
# "2" - kluso
def is_number(txt):
res = True
try:
int(txt)
except:
res = False
return res
if bold_txt in exceptions or is_number(bold_txt):
w_lst = [] # [bold_txt]
else:
w_lst = [w.strip() for w in bold_txt.split(",")]
def remove_bad_suffix(w):
for suffix in [
":", # boarding:
" 1", # can 1
]:
if w.endswith(suffix):
w = w[:-len(suffix)]
return w
# только первое слово - корень
# kost/i, ~o cost; multe~a expensive
if radix is None:
radix = w_lst[0]
slash = radix.find("/")
if slash >= 0:
radix = radix[:slash]
radix = remove_bad_suffix(radix)
for w in w_lst:
for no_tilda_pattern in [
"(aerarmea) generalo", # air
"koncerne (with accus)", # as
"~ on daŭri", # run
]:
if idx != 0 and w.find("~") == -1 and txt.find(
no_tilda_pattern) != -1:
w = "~ " + w
# :TRICKY: некоторые термины содержат " ~ ", но без
# ручного анализа правильное значение не подставишь:
# - lav/i wash tr; ~ sin get washed, wash (oneself)
# - est/i be; ~as (there) is/are; kio ~ al vi? what’s the matter? [skip]
w = w.replace("/", "").replace("~", radix)
# Kaliforni/o California; ≈o californium
change_case = w.find("≈") >= 0
if change_case:
w = w.replace("≈", radix)
# :REFACTOR:
w = w[0].swapcase() + w[1:]
# digital/o 1 digitalis, foxglove; 2 ~a img2.png digital [= cifereca]
if w.startswith("2 "):
w = w[2:]
w = remove_bad_suffix(w)
# Prote/o Proteus; ≈a protean; ≈o 1 protea (flower); 2 olm (amphibian)
# errors needs to be fixed by upstream
if w in ['a', 'o']:
continue
if w == 'la' and txt.find(
"da is not used before la, other") != -1:
continue
make_kondratjev.with_x_hdw(w, on_word)
is_first = False
assert lst
#print(lst)
on_parsed_article(lst, parse_vip.gen_html_text(p)) # txt)
if not txt:
found_empty_p = True
def main():
prefix = "/home/ilya/.stardict/dic/esperanto"
make_wells(15, 42, o_p.join(prefix, "wells-eo-en/wells-eo-en.txt"))
make_wells(44, 69, o_p.join(prefix, "wells-en-eo/wells-en-eo.txt"))
def make_fpath(fname):
return o_p.join(fdir, fname)
def setup(cont_name, keyfile, image, need_install_python=False, **kwargs):
with open(keyfile) as f:
pub_key = f.read()
# добавляем директорию ansible_plugins в список, где искать доп. расширения
# (до run_api())
from ansible import utils
import o_p
import os
get_par = os.path.dirname
utils.plugins.push_basedir(
o_p.join(get_par(get_par(__file__)), "ansible_plugins"))
import ans_module
labels = {"disvolvu": "test"}
# настраиваем docker без with_sudo=True
res = ans_module.run_api("docker",
name=cont_name,
image=image,
labels=labels,
docker_api_version="auto",
**kwargs)
assert res
client = create_dclient()
cont = find_container(cont_name, client)
assert cont
ip_addr = update_ssh_config.get_ipaddr(cont, client)
#print(ip_addr)
# удаляем предыдущий возможный fingerprint
known_hosts = os.path.expanduser("~/.ssh/known_hosts")
run_command("""ssh-keygen -f "%(known_hosts)s" -R %(ip_addr)s""" %
locals())
# добавляем SSH-публичный ключ хоста в known_hosts, чтобы
# ssh не ругался при неинтерактивном соединении
from ansible.module_utils.known_hosts import add_git_host_key
# :TRICKY: настраиваем фейковый модуль
from ansible.module_utils.basic import AnsibleModule
ans_module.setup_module_arguments(None)
fake_module = AnsibleModule({})
# :KLUDGE: сделать нормальную функцию без имитации git-адреса
add_git_host_key(fake_module, "git@%(ip_addr)s/" % locals())
#
standard_password = "******"
if need_install_python:
run_command(
"sshpass -p %(standard_password)s ssh root@%(ip_addr)s apt-get -y install python"
% locals())
# в Ubuntu пакет python уже есть, поэтому
# сразу к делу - authorized_key
run_module([
{
"host": ip_addr,
"settings": {
"ansible_ssh_user": "******",
"ansible_ssh_pass": "******"
},
},
],
"authorized_key",
'',
user="******",
key=pub_key)
update_ssh_config.main()
parser.add_argument(name, help=help)
add_arg("src_file_dir", "source file directory")
add_arg("src_basename", "source file basename (w/o extension)")
add_arg("build_dir", "build directory for make")
args = parser.parse_args()
#print args
is_eq = True
ln = min(len(args.build_dir), len(args.src_file_dir))
for i in range(ln):
if args.build_dir[i] != args.src_file_dir[i]:
eq_len = i
is_eq = False
break
if is_eq:
eq_len = ln
import o_p
path = o_p.join(args.src_file_dir[eq_len:], args.src_basename + ".o")
args = "/usr/bin/make", [path]
print("Running:", *args)
# для Qt Creator'а требуется flush, а то не покажется Running:
import sys
sys.stdout.flush()
import call_cmd
call_cmd.exec_process(*args, run_new_process=False)
def setup(cont_name, keyfile, image, need_install_python=False, **kwargs):
with open(keyfile) as f:
pub_key = f.read()
# добавляем директорию ansible_plugins в список, где искать доп. расширения
# (до run_api())
from ansible import utils
import o_p
import os
get_par = os.path.dirname
utils.plugins.push_basedir(o_p.join(get_par(get_par(__file__)), "ansible_plugins"))
import ans_module
labels = {
"disvolvu": "test"
}
# настраиваем docker без with_sudo=True
res = ans_module.run_api("docker", name=cont_name, image=image,
labels=labels, docker_api_version="auto", **kwargs)
assert res
client = create_dclient()
cont = find_container(cont_name, client)
assert cont
ip_addr = update_ssh_config.get_ipaddr(cont, client)
#print(ip_addr)
# удаляем предыдущий возможный fingerprint
known_hosts = os.path.expanduser("~/.ssh/known_hosts")
run_command("""ssh-keygen -f "%(known_hosts)s" -R %(ip_addr)s""" % locals())
# добавляем SSH-публичный ключ хоста в known_hosts, чтобы
# ssh не ругался при неинтерактивном соединении
from ansible.module_utils.known_hosts import add_git_host_key
# :TRICKY: настраиваем фейковый модуль
from ansible.module_utils.basic import AnsibleModule
ans_module.setup_module_arguments(None)
fake_module = AnsibleModule({})
# :KLUDGE: сделать нормальную функцию без имитации git-адреса
add_git_host_key(fake_module, "git@%(ip_addr)s/" % locals())
#
standard_password = "******"
if need_install_python:
run_command("sshpass -p %(standard_password)s ssh root@%(ip_addr)s apt-get -y install python" % locals())
# в Ubuntu пакет python уже есть, поэтому
# сразу к делу - authorized_key
run_module([
{
"host": ip_addr,
"settings": {
"ansible_ssh_user": "******",
"ansible_ssh_pass": "******"
},
},
], "authorized_key", '', user="******", key=pub_key)
update_ssh_config.main()
def copy_prj_fname(prj_fname, dst_fpath):
o_p.force_makedirs(os.path.dirname(dst_fpath))
shutil.copy(o_p.join(prj_fdir, prj_fname), dst_fpath)
def copy_prj_fname(prj_fname, dst_fpath):
o_p.force_makedirs(os.path.dirname(dst_fpath))
shutil.copy(o_p.join(prj_fdir, prj_fname), dst_fpath)
def open_xml_article(xml_fname):
xml_fname = o_p.join(xml_fpath, xml_fname)
tree = open_xml_tree(xml_fname)
return tree
def parse_dictionary(on_parsed_article):
dirname = os.path.dirname
prefix_eoru = dirname(dirname(dirname(__file__)))
unpacked_epub = o_p.join(prefix_eoru, "stuff/Wells/decrypted/depacked")
for num in range(first_num, last_num+1):
#num = 29 # K
src_fname = o_p.join(unpacked_epub, "OEBPS/%03d.html" % num)
body = get_html_body(src_fname, True)
found_empty_p = False
# альтернативный способ - .getchildren() + проверка .tag
for p in body.iterfind(xhtml_tag("p")):
txt = p.text_content().strip()
if found_empty_p and txt:
# очередная статья
#print(txt)
radix = None
lst = []
def on_word(word):
# <b>Kaboverd/o</b><b> </b>Cape Verde
if word:
lst.append(word)
key_elements = list(parse_vip.iter_tags(p, xhtml_tag("b")))
assert key_elements
for idx, el in enumerate(key_elements):
bold_txt = el.text_content().strip()
exceptions = [
"li diris, ~ ŝi atendas", # ke
"~e, ke", # kondiĉo
]
# "2" - kluso
def is_number(txt):
res = True
try:
int(txt)
except:
res = False
return res
if bold_txt in exceptions or is_number(bold_txt):
w_lst = [] # [bold_txt]
else:
w_lst = [w.strip() for w in bold_txt.split(",")]
def remove_bad_suffix(w):
for suffix in [
":", # boarding:
" 1", # can 1
]:
if w.endswith(suffix):
w = w[:-len(suffix)]
return w
# только первое слово - корень
# kost/i, ~o cost; multe~a expensive
if radix is None:
radix = w_lst[0]
slash = radix.find("/")
if slash >= 0:
radix = radix[:slash]
radix = remove_bad_suffix(radix)
for w in w_lst:
for no_tilda_pattern in [
"(aerarmea) generalo", # air
"koncerne (with accus)", # as
"~ on daŭri", # run
]:
if idx != 0 and w.find("~") == -1 and txt.find(no_tilda_pattern) != -1:
w = "~ " + w
# :TRICKY: некоторые термины содержат " ~ ", но без
# ручного анализа правильное значение не подставишь:
# - lav/i wash tr; ~ sin get washed, wash (oneself)
# - est/i be; ~as (there) is/are; kio ~ al vi? what’s the matter? [skip]
w = w.replace("/", "").replace("~", radix)
# Kaliforni/o California; ≈o californium
change_case = w.find("≈") >= 0
if change_case:
w = w.replace("≈", radix)
# :REFACTOR:
w = w[0].swapcase() + w[1:]
# digital/o 1 digitalis, foxglove; 2 ~a img2.png digital [= cifereca]
if w.startswith("2 "):
w = w[2:]
w = remove_bad_suffix(w)
# Prote/o Proteus; ≈a protean; ≈o 1 protea (flower); 2 olm (amphibian)
# errors needs to be fixed by upstream
if w in ['a', 'o']:
continue
if w == 'la' and txt.find("da is not used before la, other") != -1:
continue
make_kondratjev.with_x_hdw(w, on_word)
is_first = False
assert lst
#print(lst)
on_parsed_article(lst, parse_vip.gen_html_text(p)) # txt)
if not txt:
found_empty_p = True
def make_dpath(*fname):
#download_data_fdir = o_p.join(project_dir_fname, "download_data")
assert download_data_fdir, "call setup_download_data() first"
return o_p.join(download_data_fdir, *fname)
parser.add_argument(name, help=help)
add_arg("src_file_dir", "source file directory")
add_arg("src_basename", "source file basename (w/o extension)")
add_arg("build_dir", "build directory for make")
args = parser.parse_args()
#print args
is_eq = True
ln = min(len(args.build_dir), len(args.src_file_dir))
for i in range(ln):
if args.build_dir[i] != args.src_file_dir[i]:
eq_len = i
is_eq = False
break
if is_eq:
eq_len = ln
import o_p
path = o_p.join(args.src_file_dir[eq_len:], args.src_basename + ".o")
args = "/usr/bin/make", [path]
print("Running:", *args)
# для Qt Creator'а требуется flush, а то не покажется Running:
import sys
sys.stdout.flush()
import call_cmd
call_cmd.exec_process(*args, run_new_process=False)
def main():
dct_prefix = "/home/ilya/.stardict/dic/esperanto"
dst_fname = o_p.join(dct_prefix, "REVO_Eksplika/Eksplika-REVO.txt")
# :REFACTOR:
dirname = os.path.dirname
# :REFACTOR: realpath() - apliki tion al ĉiuj uzoj
prj_fdir = dirname(os.path.realpath(__file__))
import shutil
def copy_prj_fname(prj_fname, dst_fpath):
o_p.force_makedirs(os.path.dirname(dst_fpath))
shutil.copy(o_p.join(prj_fdir, prj_fname), dst_fpath)
prefix_eoru = dirname(dirname(prj_fdir))
unpacked_revo = o_p.join(prefix_eoru, "stuff/revo/revo")
dictionaries = {}
with make_gen_accumulator() as add_gen:
def create_dictionary(dst_fname, css_link=None):
remove_srcfile = True # False #
on_article = add_gen(make_kondratjev.dictionary_generator(dst_fname, css_text=None, is_html=True, remove_srcfile=remove_srcfile))
if css_link:
orig_on_article = on_article
def on_article(key_names, txt):
css_link
txt = """<link href="%(css_link)s" rel="stylesheet" type="text/css" />%(txt)s""" % locals()
return orig_on_article(key_names, txt)
return on_article
on_explika_article = create_dictionary(dst_fname, "revo.css")
res_fdir = o_p.join(dirname(dst_fname), "res")
copy_prj_fname("sample/revo/revo.css", o_p.join(res_fdir, "revo.css"))
# kopias figurojn por beleco
dst_smb = o_p.join(res_fdir, "smb")
if not o_p.exists(dst_smb):
shutil.copytree(o_p.join(unpacked_revo, "smb"), dst_smb)
xml_fpath = o_p.join(unpacked_revo, "xml")
def open_xml_article(xml_fname):
xml_fname = o_p.join(xml_fpath, xml_fname)
tree = open_xml_tree(xml_fname)
return tree
def fname2prefix(src_fname):
return o_p.without_ext(src_fname)
prefix_dct = {}
def get_words(prefix):
words = prefix_dct.get(prefix)
if words is None:
words = prefix_dct[prefix] = []
tree = open_xml_article(prefix + ".xml")
for drv, headwords in for_drv_words_headwords(tree):
words.extend(calc_words(headwords))
#print(words)
#print(rvut_definitions.get_translations(drv).get("de"))
#print()
return words
fname_lst = os.listdir(xml_fpath)
if False: # True: #
fname_lst = [
"ten.xml",
"distin.xml",
"apenau.xml", # <trd> in <subdrv>
"pri.xml", # artikolo sen <drv>
"sur.xml", # <ekz> ne en <(sub)snc>, sed en <subart>
"al.xml", # <trdgrp> ĝuste en <art>
"stift.xml", # kaj <ekz> ankaŭ en <art>
"lima.xml", # перевод относился к <kap>, хотя был внутри текста (гад с 'la') - и таких статей много
"kverk.xml", # diversaj homaj eraroj
"jxak1.xml",
"anim.xml", # <ekz> sen <ind>
"blank.xml", #
"milv.xml", # <bld> anstataŭ <>ekz
"hel.xml", # trdgrp en <dif>
"hazard.xml", # malplena trd etikedo
"iks.xml", # vortoj kun signo '|'
]
for src_fname in fname_lst:
prefix = fname2prefix(src_fname)
all_names = get_words(prefix)
html_fname = o_p.join(unpacked_revo, "art", prefix + ".html")
body = make_wells.get_html_body(html_fname, False)
h1 = body.find("h1")
hr = body.find("hr")
div = etree.Element("div")
el = h1.getnext()
while el != hr:
div.append(el)
el = h1.getnext()
def append_sub(name):
sub_el = body.find("div[@class='%(name)s']" % locals())
if not(sub_el is None):
div.append(etree.Element("hr"))
div.append(sub_el)
append_sub("fontoj")
append_sub("notoj")
# renovigas referencojn en stilo
# kapt.html#kapt.0i => bword://kapti#kapt.0i
for lnk in parse_vip.iter_tags(div, "a"):
href = lnk.get("href")
if href:
m = re.match(r"(?P<lnk_fname>[^/]+\.html)#(?P<anchor>.+)$", href)
if m:
lnk_fname, anchor = m.group("lnk_fname"), m.group("anchor")
lnk_word = get_words(fname2prefix(lnk_fname))[0]
# GD ne atentas #anchor, ColorDict - eĉ rifuzas sekvi la ligilon
#lnk.set("href", "bword://%(lnk_word)s#%(anchor)s" % locals())
lnk.set("href", "bword://%(lnk_word)s#%(anchor)s" % locals())
# :REFACTOR:
for img in parse_vip.iter_tags(div, "img"):
src = img.get("src")
if src:
# egala funkciado por Goldendict (GD) k ColorDict (CD)
m = re.match(r"^\.\./", src)
if m:
img.set("src", src[3:])
txt = parse_vip.gen_html_text(div)
#print(txt)
on_explika_article(all_names, txt)
# eo-nacia vortaro
national_dct = {}
tree = open_xml_article(src_fname)
def append_translations(translations, src_trs):
for lang, lst in src_trs.items():
translations[lang] = lst + translations.setdefault(lang, [])
used_tr_nodes = {}
national_headwords = {}
def get_count_translations(node):
res = rvut_definitions.get_translations(node)
# hazard.xml havas malplena tradukojn
clean_res = {}
for lang, lst in res.items():
lst = list(filter(bool, lst))
if lst:
clean_res[lang] = lst
res = clean_res
append_translations(national_headwords, res)
# :REFACTOR:
for trd in node.findall('trd'):
used_tr_nodes[trd] = True
for trdp in node.findall('trdgrp'):
used_tr_nodes[trdp] = True
for trd in trdp.findall('trd'):
used_tr_nodes[trd] = True
return res
def iterate_translations(translations, sub_node_dct, numerator_func, final_sep):
for lang in sub_node_dct.keys() | translations.keys():
yield lang, gen_trans_text(lang, sub_node_dct, numerator_func, translations, final_sep)
def notify_node(warning_txt, node):
print(warning_txt, src_fname, parse_vip.gen_html_text(node))
# :TRICKY: plej simpla maniero por kalkuki jam traktitajn nodojn
ekz_node_set = set()
def find_ekz_translations(ekz_dct, node, flat_translations):
#for trd in parse_vip.iter_tags(node, "ekz/trd|trdgrp"):
def trd_iter(ekz_name, name):
return parse_vip.iter_tags(node, "%(ekz_name)s/%(name)s" % locals())
def trd_iters(ekz_name):
return trd_iter(ekz_name, "trd"), trd_iter(ekz_name, "trdgrp")
for trd in itertools.chain(*(trd_iters("ekz") + trd_iters("bld"))):
ekz = trd.getparent()
if ekz in ekz_node_set:
continue
else:
ekz_node_set.add(ekz)
def make_orig_txt(ind_node):
return ', '.join(rvut_words.get_words_from_kap(ind_node))
ind_node = ekz.find('ind')
if ind_node is None:
# kalkulas orig_txt mem, kolektante ĉiujn etikedojn ĝis apero de trd aŭ trdgrp
# anim.xml:
# <ekz>
# <tld/>ita parolado<fnt>K</fnt>,
# <trd lng="hu">lelkes beszéd</trd>
# </ekz>
ind_node = etree.Element("ind")
ind_node.text = ekz.text
for child in ekz.getchildren():
if child.tag in ["trd", "trdgrp"]:
break
else:
child = copy.deepcopy(child)
ind_node.append(child)
tree.append(ind_node)
orig_txt = make_orig_txt(ind_node)
ind_node.getparent().remove(ind_node)
else:
orig_txt = make_orig_txt(ind_node)
for lang, tr_lst in get_count_translations(ekz).items():
# :REFACTOR:
lst = ekz_dct.setdefault(lang, [])
tr_lst = ", ".join(tr_lst)
ekz_txt = "<i><b>%(orig_txt)s</b>: %(tr_lst)s</i>" % locals()
lst.append(ekz_txt)
#return
# :TRICKY: iuj <trd> kumulas tradukon mem k indikon de originala nomo (Latina prezipe) =>
# nur <trd> povas esti tia, ne <trdgrp>, ĉar tio estas perokula etikedo (angla - tag)
# malĝuste - hel.xml!
rest_translations = {}
for trd in parse_vip.iter_tags(node, "trd"):
if trd not in used_tr_nodes:
par_node = trd.getparent()
if par_node.tag == "trdgrp":
lang = par_node.get("lng")
used_tr_nodes[par_node] = True
else:
lang = trd.get("lng")
foreign_word = rvut_flatten.flatten_node(trd)
if foreign_word:
# :REFACTOR:
rest_translations.setdefault(lang, []).append(foreign_word)
# :REFACTOR:
used_tr_nodes[trd] = True
append_translations(flat_translations, rest_translations)
append_translations(national_headwords, rest_translations)
def append_ekz_translations(dct, ekz_dct):
# :TRICKY: ke lasi subsnc_dct simplan kaj ne ŝanĝi iterater_translations,
# do ĵuste aldonas ekzemplojn al lasta ero de subsnc_dct
for lang, ekz_lst in ekz_dct.items():
ekz_txt = "; ".join(ekz_lst)
lst = dct.setdefault(lang, [])
if lst:
lst[-1] += "; " + ekz_txt
else:
lst.append(ekz_txt)
def append_national_article(lang, names, txt):
o_p_article, dst_fname = dictionaries.get(lang, (None, None))
if o_p_article is None:
dict_fpath = o_p.join(dct_prefix, "REVO_%(lang)s" % locals())
# :REFACTOR:
dst_fname = o_p.join(dict_fpath, "REVO-%(lang)s.txt" % locals())
o_p_article = create_dictionary(dst_fname, "revo-traduko.css")
dictionaries[lang] = o_p_article, dst_fname
copy_prj_fname("sample/revo/eo-nacia/revo-traduko.css", o_p.join(dict_fpath, "res/revo-traduko.css"))
o_p_article(names, txt)
def append_row(translations, snc_dct, headwords, drv):
# sur.xml: <ekz> povas esti ekster ajna <snc>
ekz_dct = {}
find_ekz_translations(ekz_dct, drv, translations)
append_ekz_translations(translations, ekz_dct)
assert headwords
hw_txt = "<b>%s</b>" % "</b>, <b>".join(headwords)
typ = None
vspec = drv.find("gra/vspec")
if vspec is not None:
typ = vspec.text
if typ:
hw_txt = "%(hw_txt)s <i>%(typ)s</i>" % locals()
for lang, tr_txt in iterate_translations(translations, snc_dct, arab_num, " <b>|</b> "):
opa_args = national_dct.setdefault(lang, ([], []))
names, txt = opa_args
names.extend(calc_words(headwords))
row_txt = """<div class="paragrafo">%(hw_txt)s %(tr_txt)s</div>""" % locals()
txt.append(row_txt)
# nacia-eo article
n_keywords = national_headwords.get(lang)
assert n_keywords
# devas purigi poste originalan n_keywords, ne ŝanĝitan
#n_keywords = [word.replace("|", "/") for word in n_keywords]
clean_keywords = [word.replace("|", "/") for word in n_keywords]
append_national_article(lang, clean_keywords, row_txt)
n_keywords.clear()
for drv, headwords in for_drv_words_headwords(tree):
#print(src_fname)
#print(translations)
#print()
def latin_num(i):
return "%(chr(ord('a') + i))s)" % s_.EvalFormat()
snc_dct = {}
ekz_snc_dct = {}
for snc in parse_vip.iter_tags(drv, "snc"):
subsnc_dct = {}
ekz_subsnc_dct = {}
for subsnc in parse_vip.iter_tags(snc, "subsnc"):
subsnc_translations = get_count_translations(subsnc)
for lang, tr_lst in subsnc_translations.items():
lst = subsnc_dct.setdefault(lang, [])
lst.append(", ".join(tr_lst))
find_ekz_translations(ekz_subsnc_dct, subsnc, subsnc_dct)
append_ekz_translations(subsnc_dct, ekz_subsnc_dct)
for lang, tr_txt in iterate_translations(get_count_translations(snc), subsnc_dct, latin_num, "; "):
lst = snc_dct.setdefault(lang, [])
lst.append(tr_txt)
find_ekz_translations(ekz_snc_dct, snc, snc_dct)
def arab_num(i):
return "<b>%(i+1)s.</b>" % s_.EvalFormat()
append_ekz_translations(snc_dct, ekz_snc_dct)
def merge_trs(translations, drv):
src_trs = get_count_translations(drv)
append_translations(translations, src_trs)
# ankoraŭ estas iome da <subdrv> en <drv> => aldonu
translations = {}
for subdrv in parse_vip.iter_tags(drv, "subdrv"):
merge_trs(translations, subdrv)
# k subart ankaŭ eblas havi <trd> ĝuste en subart - sur.xml: hu => rá-
merge_trs(translations, drv)
append_row(translations, snc_dct, headwords, drv)
# :TRICKY: al.xml havas tradukojn ekstere subart, drv
art_node = tree.find("art")
append_row(get_count_translations(art_node), {}, find_kap_words(art_node), art_node)
for lang, opa_args in national_dct.items():
names, txt = opa_args
append_national_article(lang, names, "".join(txt))
strict_check = False # True #
def alarm_not_processed(trd):
is_ok = trd in used_tr_nodes
if not is_ok:
if strict_check:
assert is_ok
else:
notify_node("Not processed trd:", trd.getparent())
# kontrolo, ke ĉiuj nodoj estas traktita
for trd in parse_vip.iter_tags(tree, "trd"):
alarm_not_processed(trd)
for trd in parse_vip.iter_tags(tree, "trdgrp"):
alarm_not_processed(trd)
# zip'u vortarojn
revo_dicts_fpath = o_p.join(dirname(unpacked_revo), "revo-dicts")
o_p.force_makedirs(revo_dicts_fpath)
# shutil povas zipfile!
#import zipfile
print("\nAtingeblaj REVO vortaroj:")
def zip_dict(dst_fname):
dir_fpath, basename = os.path.split(dst_fname)
root_dir, dir_fname = os.path.split(dir_fpath)
# uzado de zip ne funkciigas, rompas Colordict - t.e. la lasta komencas blinki senfine
#fmt = "zip"
fmt = "gztar"
# se vortara datumo estas ĝuste en arkivo, ne en ia dosierujo, do CSS/figuroj estas ne trovataj
save_without_folder = False # True #
if save_without_folder:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname), fmt, dir_fpath)
else:
fname = shutil.make_archive(o_p.join(revo_dicts_fpath, dir_fname), fmt, root_dir, base_dir=dir_fname)
ifo_fname = os.path.splitext(dst_fname)[0] + ".ifo"
with open(ifo_fname) as ifo_f:
properties = {}
for line in ifo_f:
lst = line.split("=")
if len(lst) >= 2:
key, value = lst[0].strip(), lst[1].strip()
if key and value:
properties[key] = value
words_cnt = int(properties.get("wordcount"))
synwordcount = properties.get("synwordcount")
if synwordcount:
words_cnt += int(synwordcount)
fname = os.path.basename(fname)
# du spacetoj fine estas por Markdown liniavanco
print("http://new.bombono.org/download/revo/%(fname)s\t%(words_cnt)s " % locals())
zip_dict(dst_fname)
for lang, (func, dst_fname) in dictionaries.items():
zip_dict(dst_fname)
