def _output_tsv(self, output_file):
"""
Output to TSV
"""
for fragment in self.fragments:
text = fragment.text_fragment
output_file.write("%s\t%s\t%s\n" % (
gf.time_to_ssmmm(fragment.begin),
gf.time_to_ssmmm(fragment.end),
text.identifier
))
def format(self, syncmap):
msg = []
msg.append(u"<?xml version=\"1.0\" encoding=\"UTF-8\" ?>")
msg.append(u"<map>")
for fragment in syncmap.fragments:
msg.append(u" <fragment>")
msg.append(u" <identifier>%s</identifier>" % fragment.text_fragment.identifier)
msg.append(u" <start>%s</start>" % gf.time_to_ssmmm(fragment.begin))
msg.append(u" <end>%s</end>" % gf.time_to_ssmmm(fragment.end))
msg.append(u" </fragment>")
msg.append(u"</map>")
return u"\n".join(msg)
def _output_txt(self, output_file):
"""
Output to TXT
"""
for fragment in self.fragments:
text = fragment.text_fragment
output_file.write("%s %s %s \"%s\"\n" % (
text.identifier,
gf.time_to_ssmmm(fragment.begin),
gf.time_to_ssmmm(fragment.end),
text.text
))
def visit_children(node, parent_elem):
""" Recursively visit the fragments_tree """
for child in node.children_not_empty:
fragment = child.value
fragment_elem = etree.SubElement(parent_elem, "fragment")
fragment_elem.attrib["id"] = fragment.text_fragment.identifier
fragment_elem.attrib["begin"] = gf.time_to_ssmmm(fragment.begin)
fragment_elem.attrib["end"] = gf.time_to_ssmmm(fragment.end)
for line in fragment.text_fragment.lines:
line_elem = etree.SubElement(fragment_elem, "line")
line_elem.text = line
children_elem = etree.SubElement(fragment_elem, "children")
visit_children(child, children_elem)
def format(self, syncmap):
msg = []
msg.append(u"<?xml version=\"1.0\" encoding=\"UTF-8\" ?>")
msg.append(u"<map>")
for fragment in syncmap.fragments:
msg.append(u" <fragment>")
msg.append(u" <identifier>%s</identifier>" %
fragment.text_fragment.identifier)
msg.append(u" <start>%s</start>" %
gf.time_to_ssmmm(fragment.begin))
msg.append(u" <end>%s</end>" % gf.time_to_ssmmm(fragment.end))
msg.append(u" </fragment>")
msg.append(u"</map>")
return u"\n".join(msg)
def _write_xml_legacy(self, output_file):
"""
Write to XML file (legacy format)
"""
output_file.write("<?xml version=\"1.0\" encoding=\"UTF-8\" ?>\n")
output_file.write("<map>\n")
for fragment in self.fragments:
text = fragment.text_fragment
output_file.write(" <fragment>\n")
output_file.write(" <identifier>%s</identifier>\n" % text.identifier)
output_file.write(" <start>%s</start>\n" % gf.time_to_ssmmm(fragment.begin))
output_file.write(" <end>%s</end>\n" % gf.time_to_ssmmm(fragment.end))
output_file.write(" </fragment>\n")
output_file.write("</map>")
def visit_children(node):
""" Recursively visit the fragments_tree """
output_fragments = []
for child in node.children_not_empty:
fragment = child.value
text = fragment.text_fragment
output_fragments.append({
"id": text.identifier,
"language": text.language,
"lines": text.lines,
"begin": gf.time_to_ssmmm(fragment.begin),
"end": gf.time_to_ssmmm(fragment.end),
"children": visit_children(child)
})
return output_fragments
def _write_xml(self, output_file):
"""
Write to XML file
"""
map_elem = etree.Element("map")
for fragment in self.fragments:
text = fragment.text_fragment
fragment_elem = etree.SubElement(map_elem, "fragment")
fragment_elem.attrib["id"] = text.identifier
fragment_elem.attrib["begin"] = gf.time_to_ssmmm(fragment.begin)
fragment_elem.attrib["end"] = gf.time_to_ssmmm(fragment.end)
for line in text.lines:
line_elem = etree.SubElement(fragment_elem, "line")
line_elem.text = line
self._write_tree_to_file(map_elem, output_file)
def _write_json(self, output_file):
"""
Write to JSON file
"""
output_fragments = []
for fragment in self.fragments:
text = fragment.text_fragment
output_fragment = {}
output_fragment["id"] = text.identifier
output_fragment["language"] = text.language
output_fragment["lines"] = text.lines
output_fragment["begin"] = gf.time_to_ssmmm(fragment.begin)
output_fragment["end"] = gf.time_to_ssmmm(fragment.end)
output_fragments.append(output_fragment)
output_dict = {"fragments": output_fragments}
output_file.write(json.dumps(output_dict, indent=1, sort_keys=True))
def test_time_to_ssmm(self):
tests = [
[None, "0.000"],
[0, "0.000"],
[1, "1.000"],
[1.234, "1.234"],
]
for test in tests:
self.assertEqual(gf.time_to_ssmmm(test[0]), test[1])
def test_time_to_ssmm(self):
tests = [
(None, "0.000"),
(0, "0.000"),
(1, "1.000"),
(1.234, "1.234"),
]
for test in tests:
self.assertEqual(gf.time_to_ssmmm(test[0]), test[1])
def format(self, syncmap):
smil_data = []
smil_ids = []
for fragment in syncmap.fragments:
text = fragment.text_fragment
smil_data.append({
"id": text.identifier,
"begin": gf.time_to_ssmmm(fragment.begin),
"end": gf.time_to_ssmmm(fragment.end)
})
smil_ids.append(text.identifier)
return gf.safe_unicode(
json.dumps(obj={
"smil_ids": smil_ids,
"smil_data": smil_data
},
indent=1,
sort_keys=True))
def test_time_to_ssmm(self):
tests = [
[None, "0.000"],
[0, "0.000"],
[1, "1.000"],
[1.234, "1.234"],
]
for test in tests:
self.assertEqual(gf.time_to_ssmmm(test[0]), test[1])
def test_time_to_ssmm(self):
tests = [
(None, "0.000"),
(0, "0.000"),
(1, "1.000"),
(1.234, "1.234"),
]
for test in tests:
self.assertEqual(gf.time_to_ssmmm(test[0]), test[1])
def _write_rbse(self, output_file):
"""
Write to RBSE file
"""
output_dict = {}
smil_data = []
smil_ids = []
for fragment in self.fragments:
text = fragment.text_fragment
output_fragment = {}
output_fragment["id"] = text.identifier
output_fragment["begin"] = gf.time_to_ssmmm(fragment.begin)
output_fragment["end"] = gf.time_to_ssmmm(fragment.end)
smil_ids.append(text.identifier)
smil_data.append(output_fragment)
output_dict = {
"smil_ids": smil_ids,
"smil_data": smil_data
}
output_file.write(json.dumps(output_dict, indent=1, sort_keys=True))
def visit_children(node):
""" Recursively visit the fragments_tree """
output_fragments = []
for child in node.children_not_empty:
fragment = child.value
text = fragment.text_fragment
output_fragments.append({
"id":
text.identifier,
"language":
text.language,
"lines":
text.lines,
"begin":
gf.time_to_ssmmm(fragment.begin),
"end":
gf.time_to_ssmmm(fragment.end),
"children":
visit_children(child)
})
return output_fragments
def format(self, syncmap):
smil_data = []
smil_ids = []
for fragment in syncmap.fragments:
text = fragment.text_fragment
smil_data.append({
"id": text.identifier,
"begin": gf.time_to_ssmmm(fragment.begin),
"end": gf.time_to_ssmmm(fragment.end)
})
smil_ids.append(text.identifier)
return gf.safe_unicode(
json.dumps(
obj={
"smil_ids": smil_ids,
"smil_data": smil_data
},
indent=1,
sort_keys=True
)
)
def _output_json(self, output_file):
"""
Output to JSON
"""
output_file.write("{\n")
output_file.write(" \"smil_ids\": [\n")
string = ",\n".join([" \"%s\"" % (
f.text_fragment.identifier
) for f in self.fragments])
output_file.write(string)
output_file.write("\n")
output_file.write(" ],\n")
output_file.write(" \"smil_data\": [\n")
string = ",\n".join([" { \"id\": \"%s\", \"begin\": %s, \"end\": %s }" % (
f.text_fragment.identifier,
gf.time_to_ssmmm(f.begin),
gf.time_to_ssmmm(f.end),
) for f in self.fragments])
output_file.write(string)
output_file.write("\n")
output_file.write(" ]\n")
output_file.write("}\n")
def _output_ttml(self, output_file, parameters=None):
"""
Output to TTML
"""
output_file.write("<?xml version=\"1.0\" encoding=\"UTF-8\" ?>\n")
output_file.write("<tt xmlns=\"http://www.w3.org/ns/ttml\">\n")
# TODO add metadata from parameters here?
# output_file.write(" <head/>\n")
output_file.write(" <body>\n")
output_file.write(" <div>\n")
for fragment in self.fragments:
text = fragment.text_fragment
output_file.write(" <p xml:id=\"%s\" begin=\"%s\" end=\"%s\">\n" % (
text.identifier,
gf.time_to_ssmmm(fragment.begin),
gf.time_to_ssmmm(fragment.end)
))
output_file.write(" %s\n" % text.text)
output_file.write(" </p>\n")
output_file.write(" </div>\n")
output_file.write(" </body>\n")
output_file.write("</tt>")
