def maybe_cache(s, directory):
if not s: return # in case poetry has some 2nd-language only
if '!synth:' in s and "_" in s:
textToSynth, langToSynth = gradint.textof(s), gradint.languageof(s)
elif s.endswith(gradint.extsep + "txt"):
textToSynth, langToSynth = gradint.readText(
directory + os.sep + s), gradint.languageof(
s, directory == gradint.promptsDirectory)
else:
return
if not langToSynth in langs: return # we're not caching that language
if synth_fileExists(
(textToSynth + "_" + langToSynth +
gradint.dotwav).lower()) or synth_fileExists(
(textToSynth + "_" + langToSynth + gradint.dotmp3).lower()):
return # it's already been done
if synth_fileExists(("__rejected_" + textToSynth + "_" + langToSynth +
gradint.dotwav).lower()) or synth_fileExists(
("__rejected_" + textToSynth + "_" + langToSynth +
gradint.dotmp3).lower()):
return # it's been rejected by the GUI
if (textToSynth.lower(), langToSynth) in generating:
return # we're already generating it
generating[(textToSynth.lower(), langToSynth)] = 1
sys.stderr.write("Generating %s\n" % (textToSynth, ))
toMove.append((gradint.synth_event(langToSynth, textToSynth).getSound(),
(textToSynth + "_" + langToSynth + gradint.dotwav).lower()))
def transLineToTuple(l):
x, y = l.split(None, 1)
y = "!synth:" + y
if gradint.extsep in x:
x = x[:x.rindex(
gradint.extsep
)] # make comparison easier (extensions don't matter to us)
return (gradint.textof(y) + "_" + gradint.languageof(y), x)
def link(l):
if type(l) == type([]): return link(l[-1])
if l.lower().endswith(gradint.dottxt):
l = "!synth:" + gradint.u8strip(
gradint.read(gradint.samplesDirectory + os.sep +
checkVariant(l, 1))).strip(
gradint.wsp) + "_" + gradint.languageof(l)
if "!synth:" in l:
if gradint.languageof(l) not in [
gradint.firstLanguage, gradint.secondLanguage
]:
l = gradint.textof(l) + " (" + gradint.languageof(l) + ")"
else:
l = gradint.textof(l)
return l.replace('&', '&').replace('<', '<')
if samples_url:
return '<A HREF="' + samples_url + checkVariant(l) + '"' + h5aCode(
checkVariant(l)) + '>' + wrappable(l) + '</A>'
return wrappable(l).replace('&', '&').replace('<', '<')
def getTxtLang(s):
if '!synth:' in s and "_" in s:
return gradint.textof(s).decode('utf-8'), gradint.languageof(s)
elif s.endswith(gradint.extsep + "txt"):
langToSynth = gradint.languageof(s)
if langToSynth == languageToCache:
return gradint.readText(s).decode(
'utf-8'
), langToSynth # else don't bother reading the file (it might be over ftpfs)
return None, None
def mark_as_used(s,directory):
if s and '!synth:' in s and "_" in s: textToSynth, langToSynth = gradint.textof(s),gradint.languageof(s)
elif s and s.endswith(gradint.extsep+"txt"): textToSynth, langToSynth = gradint.readText(directory+os.sep+s), gradint.languageof(s,directory==gradint.promptsDirectory)
else: return
k=(textToSynth+"_"+langToSynth).lower()
k=trans.get(k,k)
for p in ["__rejected_","__rejected__"]:
if p+k in unused: del unused[p+k]
if not k in unused:
if "_"+k in unused: k="_"+k
else: return
if also_remove_words_that_can_be_synthesized_by_partials and gradint.synth_from_partials(textToSynth,langToSynth): return
del unused[k]
def maybe_list(s, directory):
if not s: return # in case poetry has some 2nd-language only
if '!synth:' in s and "_" in s:
textToSynth, langToSynth = gradint.textof(s), gradint.languageof(s)
elif s.endswith(gradint.dottxt):
textToSynth, langToSynth = gradint.readText(
directory + os.sep + s), gradint.languageof(
s, directory == gradint.promptsDirectory)
else:
return
if not langToSynth == lang: return # we're not listing that language
if textToSynth.lower() in listed: return
d = textToSynth.decode('utf-8')
if hanzi_only and langToSynth == "zh" and not gradint.fix_compatibility(
d).replace(" ", "") == gradint.hanzi_and_punc(d):
return
global wroteChars
if paragraph_size and wroteChars and wroteChars + len(d) > paragraph_size:
toList.append("")
wroteChars = 0
wroteChars += (len(d) + 2) # comma and \n
toList.append(textToSynth + ",")
listed[textToSynth.lower()] = 1
== gradint.secondLanguage, pLines):
continue
plines2 = []
for p in pLines:
idx = line2index[p]
doneAlready[idx] = 1
prompt = d.unavail[idx][1]
equals = ""
if type(prompt) == type([]):
if len(prompt) == 3: equals = prompt[1]
elif not plines2 and not prompt == p: equals = prompt # if 1st line
if equals:
assert equals.startswith("!synth:") and gradint.languageof(
equals
) == gradint.firstLanguage, "recovery of poems with non-L1 secondary prompts not yet supported"
equals = "=" + gradint.textof(equals)
plines2.append(gradint.textof(p) + equals)
output.append((d.unavail[line2index[pLines[0]]][0], gradint.secondLanguage,
gradint.firstLanguage,
"\n".join(["begin poetry"] + plines2 + ["end poetry"])))
for count, (num, L1, L2) in zip(xrange(len(d.unavail)), d.unavail):
if count in doneAlready: continue
if type(L1) == type(L2) == type("") and L1.startswith(
"!synth:") and L2.startswith("!synth:"):
lang1, lang2 = gradint.languageof(L1), gradint.languageof(L2)
output.append((num, lang2, lang1,
"%s=%s" % (gradint.textof(L2), gradint.textof(L1))))
output.sort()
output.reverse()