def v(str):
c = iphon_configparser.parser()
c.read_string(string=str)
return commit.validate(c)
if args.bibkey:
ref_info.update(from_bibkey(args.glottocode, args.bibkey))
if args.sil_pacific:
ref_info.update(from_sil_pacific(args.sil_pacific))
if args.phon_invs_tb:
args.simple = True
args.bibkey = 'hh:hld:Namkung:Tibeto-Burman'
ref_info.update(from_bibkey(args.glottocode, args.bibkey))
ref_info['url'] = 'https://stedt.berkeley.edu/pubs_and_prods/STEDT_Monograph3_Phonological-Inv-TB.pdf'
ref_info['author'] = 'Namkung, Ju'
ini_path = find_path(args.glottocode)
# build the file
ini = iphon_configparser.parser()
ini['core'] = {
'name': args.name
, 'glottocode': args.glottocode
, 'dialect': 'OPTIONAL'
, 'dialect_name': args.dialect_name
}
ini['source'] = {
'glottolog': (args.bibkey if args.bibkey else "IDEAL")
, 'url': maybe(ref_info, 'url', "IDEAL")
, 'doi': maybe(ref_info, 'doi', "OPTIONAL")
, 'author': maybe(ref_info, 'author', "OPTIONAL (but REQUIRED if there's no glottolog ID)")
, 'title': maybe(ref_info, 'title', "OPTIONAL (but REQUIRED if there's no glottolog ID)")
, 'publisher': maybe(ref_info, 'publisher', "OPTIONAL")
, 'volume': maybe(ref_info, 'volume', "OPTIONAL")
if len(phonemes) == 0 or phonemes[0] == 'required':
raise MissingPropertyError('No phonemes given')
if 'y' in phonemes:
print('Warning: /y/ listed in phonemes - make sure you don\'t mean /j/!')
canonical_phonemes = [get_canonical(phoneme) for phoneme in phonemes]
# -- Allophonic rules tests --
for rule_raw in doculect['allophonic_rules']:
rule = parse_allophonic_rule(rule_raw)
for phoneme in rule['phonemes']:
if phoneme not in canonical_phonemes:
raise InvalidPropertyError('Phoneme {} in rule {} not listed as canonical in phonemes section'.format(phoneme, rule))
return True
if __name__ == '__main__':
filename = sys.argv[1]
doculect = iphon_configparser.parser()
file_path = path.join('doculects', '{}.ini'.format(filename))
if not(path.isdir('doculects')):
raise NotFoundError('Doculects directory not found - this script must be run from the main IPHON directory')
if not(path.isfile(file_path)):
raise NotFoundError('File not found')
doculect.read(file_path, encoding='utf-8')
validate(doculect) # if it's invalid, this will throw an exception
subprocess.run(['git', 'add', file_path])
def read_ini(path, sql):
ini = iphon_configparser.parser()
ini.read(path, encoding='utf-8')
validate(ini)
doculect = OrderedDict()
doculect['inventory_id'] = os.path.split(
path)[-1][:-4] # get the filename minus the .ini
doculect[DOCULECT_NAME_COL] = ini['core']['name']
doculect['glottocode'] = ini['core']['glottocode']
doculect['dialect'] = maybe(ini['core'], 'dialect')
doculect['dialect_name'] = maybe(ini['core'], 'dialect_name')
doculect['notes'] = '\n'.join(maybe(ini, 'notes', []))
if doculect['notes'] == '':
doculect['notes'] = None
doculect['source_bibkey'] = maybe(ini['source'],
'glottolog',
filters=INI_DEFAULTS)
doculect['source_url'] = maybe(ini['source'], 'url', filters=INI_DEFAULTS)
doculect['source_author'] = maybe(ini['source'],
'author',
filters=INI_DEFAULTS)
doculect['source_title'] = maybe(ini['source'],
'title',
filters=INI_DEFAULTS)
doculect['source_publisher'] = maybe(ini['source'],
'publisher',
filters=INI_DEFAULTS)
doculect['source_volume'] = maybe(ini['source'],
'volume',
filters=INI_DEFAULTS)
doculect['source_number'] = maybe(ini['source'],
'number',
filters=INI_DEFAULTS)
doculect['source_year'] = maybe(ini['source'],
'year',
filters=INI_DEFAULTS)
doculect['source_pages'] = maybe(ini['source'],
'pages',
filters=INI_DEFAULTS)
doculect['source_doi'] = maybe(ini['source'], 'doi', filters=INI_DEFAULTS)
if doculect['source_year'] == 'Unknown':
doculect['source_year'] = None
phonemes = list(ini['phonemes'])
allophonic_rules = [
parse_allophonic_rule(a) for a in ini['allophonic_rules']
]
# Start writing - first the doculect...
insert('doculects', doculect, return_id=True, sql=sql)
doculect_id = sql.fetchone()[0]
# ...then the language, if necessary...
language_id = find_or_create_language(doculect['glottocode'], sql=sql)
# ...then the segments...
# We'll handle featuralization later.
# We'll also figure out how to store alternate forms later;
# for now, we'll mostly mirror PSMITH's db structure,
# and store non-canonical forms (that aren't listed as allophones) as conditionless allophonic rules.
# (The eventual goal is to have everything in Haskell, so this is good enough for v0.1,
# but it's not strictly correct.)
# Another thing we'll figure out later is phoneme junctions in allophonic rules -
# e.g. s+i > s̩ / unstressed.
# These will be stored as:
# s > s̩ / unstressed (_+i)
# i > s̩ / unstressed (s+_)
# Which also isn't strictly correct, but isn't good enough.
# Rules with + in the output are just ignored, because I don't know how they should be handled, and see above re: deadlines.
# Maybe we want NoSQL for the final DB.
doculect_segment_ids = {}
for phoneme_txt in phonemes:
phoneme = parse_phoneme(phoneme_txt)
canonical_form_id = find_or_create_segment(phoneme['canonical_form'],
sql=sql)
noncanonical_forms = phoneme['noncanonical_forms']
# OK, let's just do this as string processing to ensure consistency
form_rule_str = '{} > {} / (non-canonical form - i.e. alternate representation of the phoneme)'
form_rules = [
parse_allophonic_rule(
form_rule_str.format(phoneme['canonical_form'],
noncanonical_form))
for noncanonical_form in noncanonical_forms
]
allophonic_rules += form_rules
# Then build the doculect_segments, and save them in a mapping of phoneme -> doculect_segment_id for this doculect.
insert(DOC_SEG_JOIN_TBL,
OrderedDict({
'doculect_id': doculect_id,
'segment_id': canonical_form_id,
'marginal': phoneme['marginal'],
'loan': phoneme['loan']
}),
return_id=True,
sql=sql)
doculect_segment_ids[phoneme['canonical_form']] = sql.fetchone()[0]
# ...then the allophones...
for allophonic_rule in allophonic_rules:
serialized_rules = []
# Note that 'variation' refers to whether the rule itself is obligatory or optional,
# not whether there's variance in outputs (all of which are distinct from the input) of the rule.
# That is, the presence of variation means the rule may output its input,
# and the absence of variation means it may not.
# That's the ideal, anyway. What +variation means here *in practice* is the presence of a ~ right after the >.
for allophone in allophonic_rule['allophones']:
# Skip anything with + in the output because we can't handle that in v0.1
if '+' in allophone:
continue
allophone_segment_id = find_or_create_segment(allophone, sql=sql)
# If it's a compound, store the whole compound as text, I guess.
compound = None
if len(allophonic_rule['phonemes']) > 1:
compound = '+'.join(allophonic_rule['phonemes'])
for phoneme in allophonic_rule['phonemes']:
rule = OrderedDict({
'doculect_segment_id':
doculect_segment_ids[phoneme],
'allophone_id':
allophone_segment_id,
'variation':
allophonic_rule['rule_type'] == 'variant',
'compound':
'+'.join(allophonic_rule['phonemes'])
if len(allophonic_rule['phonemes']) > 1 else None,
'environment':
allophonic_rule['environment']
})
insert('allophones', rule, sql=sql)