def __init__(self, lexicon_name):
super(LexiconTrie, self).__init__(self.dict_letter_scope)
for entry in LexiconEntrySet({"parent_lexicon": lexicon_name}, sort=[("_id", -1)]):
self[hebrew.strip_nikkud(entry.headword)] = entry.headword
for ahw in getattr(entry, "alt_headwords", []):
self[hebrew.strip_nikkud(ahw)] = entry.headword
def clean_line(line):
line = strip_nikkud(line)
line = re.sub(u':', '', line)
reg_parentheses = re.compile(u'\((.*?)\)')
reg_brackets = re.compile(u'\[(.*?)\]')
in_per = reg_parentheses.search(line)
in_bra = reg_brackets.search(line)
reg_ayyen_tur = re.compile(u'''ו?(עיין|עי'|ע"ש) בטור''')
line = re.sub(u'\[.*?אלפס.*?\]', u'', line)
line = re.sub(u'טור ו?שו"ע', u'טוש"ע', line)
pos = re.search(reg_ayyen_tur, line)
if pos:
line = line[:pos.start()]
if in_per:
if in_bra:
clean = re.sub(reg_brackets, ur'\1', line) # brackets are always correct
clean = re.sub(reg_parentheses, '', clean)
else:
clean = re.sub(reg_parentheses, ur'\1', line)
elif in_bra:
clean = re.sub(reg_brackets, ur'\1', line) # brackets are always correct
else:
clean = line
return clean
def clean_line(line):
line = strip_nikkud(line)
replace_dict = {u'[.:\?]': u'', u'[”״]': u'"', u'[’׳]': u"'"} #note put \. in the file/ how can i check if it is right?
line = multiple_replace(line, replace_dict, using_regex=True)
# line = re.sub(u'[:\?]', '', line)
# line = re.sub(u'”', u'"', line)
reg_parentheses = re.compile(u'\((.*?)\)')
reg_brackets = re.compile(u'\[(.*?)\]')
in_per = reg_parentheses.search(line)
in_bra = reg_brackets.search(line)
reg_ayyen_tur = re.compile(u'''ו?(עיין|עי'|ע"ש) בטור''')
reg_lo_manu = re.compile(u'''(?P<a>(\u05d0\u05da )?\u05dc\u05d0 \u05de\u05e0(.*?))(\u05e1\u05de"?\u05d2|\u05e8\u05de\u05d1"?\u05dd|\u05d8\u05d5\u05e8|\n)''')
line = re.sub(u'\[.*?אלפס.*?\]', u'', line)
line = re.sub(u'טור ו?שו"ע', u'טוש"ע', line)
f_ayyen = re.search(reg_ayyen_tur, line)
f_lo_manu = re.search(reg_lo_manu, line)
if f_ayyen:
line = line[:f_ayyen.start()]
if f_lo_manu:
line = re.sub(f_lo_manu.group('a'), u"", line)
if in_per:
if in_bra:
clean = re.sub(reg_brackets, ur'\1', line) # brackets are always correct
clean = re.sub(reg_parentheses, '', clean)
else:
clean = re.sub(reg_parentheses, ur'\1', line)
elif in_bra:
clean = re.sub(reg_brackets, ur'\1', line) # brackets are always correct
else:
clean = line
return clean
def clean_line(line):
line = strip_nikkud(line)
replace_dict = {u'[.:\?]': u'', u'[”״]': u'"', u'[’׳]': u"'"} #note put \. in the file/ how can i check if it is right?
line = multiple_replace(line, replace_dict, using_regex=True)
# line = re.sub(u'[:\?]', '', line)
# line = re.sub(u'”', u'"', line)
reg_parentheses = re.compile(u'\((.*?)\)')
reg_brackets = re.compile(u'\[(.*?)\]')
in_per = reg_parentheses.search(line)
in_bra = reg_brackets.search(line)
reg_ayyen_tur = re.compile(u'''ו?(עיין|עי'|ע"ש) בטור''')
reg_lo_manu = re.compile(u'''(?P<a>(\u05d0\u05da )?\u05dc\u05d0 \u05de\u05e0(.*?))(\u05e1\u05de"?\u05d2|\u05e8\u05de\u05d1"?\u05dd|\u05d8\u05d5\u05e8|\n)''')
line = re.sub(u'\[.*?אלפס.*?\]', u'', line)
line = re.sub(u'טור ו?שו"ע', u'טוש"ע', line)
f_ayyen = re.search(reg_ayyen_tur, line)
f_lo_manu = re.search(reg_lo_manu, line)
if f_ayyen:
line = line[:f_ayyen.start()]
if f_lo_manu:
line = re.sub(f_lo_manu.group('a'), u"", line)
if in_per:
if in_bra:
clean = re.sub(reg_brackets, ur'\1', line) # brackets are always correct
clean = re.sub(reg_parentheses, '', clean)
else:
clean = re.sub(reg_parentheses, ur'\1', line)
elif in_bra:
clean = re.sub(reg_brackets, ur'\1', line) # brackets are always correct
else:
clean = line
return clean
def clean_line(line):
line = strip_nikkud(line)
line = re.sub(u':', '', line)
reg_parentheses = re.compile(u'\((.*?)\)')
reg_brackets = re.compile(u'\[(.*?)\]')
in_per = reg_parentheses.search(line)
in_bra = reg_brackets.search(line)
reg_ayyen_tur = re.compile(u'''ו?(עיין|עי'|ע"ש) בטור''')
line = re.sub(u'\[.*?אלפס.*?\]', u'', line)
line = re.sub(u'טור ו?שו"ע', u'טוש"ע', line)
pos = re.search(reg_ayyen_tur, line)
if pos:
line = line[:pos.start()]
if in_per:
if in_bra:
clean = re.sub(reg_brackets, ur'\1',
line) # brackets are always correct
clean = re.sub(reg_parentheses, '', clean)
else:
clean = re.sub(reg_parentheses, ur'\1', line)
elif in_bra:
clean = re.sub(reg_brackets, ur'\1',
line) # brackets are always correct
else:
clean = line
return clean
def body(self):
self.load_toc()
self.click_toc_category("Midrash")
self.click_toc_text("Ein Yaakov")
self.click_source_title()
self.click_masechet_and_chapter('2', '3')
section = self.get_section_txt('1')
assert 'רבי זירא הוה קא משתמיט' in strip_nikkud(section)
def find_in_segment(self, st, lang='he', citing_only=False, replace=True):
#todo: implemant replace = True
"""
Returns a list of Ref objects derived from string
:param string st: the input string
:param lang: "he" note: "en" is not yet supported in ibid
:param citing_only: boolean whether to use only records explicitly marked as being referenced in text.
:return: list of :class:`Ref` objects, list of locations and list of ref types (either REF or SHAM, defined in CitationFinder)
"""
refs = []
locations = []
types = []
failed_refs = []
# failed_non_refs = []
failed_shams = []
assert lang == 'he'
# todo: support english
st = strip_nikkud(st)
all_refs = self._citationFinder.get_potential_refs(
st, lang, citing_only=citing_only)
for item, location, type in all_refs:
if type == CitationFinder.REF_INT:
try:
refs += [
self._tr.resolve(item.index_node.full_title(),
item.sections)
]
locations += [location]
types += [type]
#refs += [(self._tr.resolve(item.index_node.full_title(), item.sections), 'ref')]
except (IbidRefException, IbidKeyNotFoundException) as e:
failed_refs += [item.normal()]
elif type == CitationFinder.NON_REF_INT or type == CitationFinder.IGNORE_INT:
# failed_non_refs += [item.group()]
self._tr.ignore_book_name_keys()
elif type == CitationFinder.SHAM_INT:
try:
if isinstance(item, str):
refs += [self._tr.resolve(None, match_str=item)]
locations += [location]
types += [type]
#refs += [(self._tr.resolve(None, match_str=item), 'sham')]
else:
refs += [self._tr.resolve(item[0], sections=item[1])]
locations += [location]
types += [type]
#refs += [(self._tr.resolve(item[0], sections=item[1]), 'sham')]
except (IbidRefException, IbidKeyNotFoundException,
InputError) as e:
failed_shams += [item]
return refs, locations, types # , failed_refs, failed_non_refs, failed_shams
def find_in_segment(self, st, lang='he', citing_only=False, replace=True):
#todo: implemant replace = True
"""
Returns a list of Ref objects derived from string
:param string st: the input string
:param lang: "he" note: "en" is not yet supported in ibid
:param citing_only: boolean whether to use only records explicitly marked as being referenced in text.
:return: list of :class:`Ref` objects, list of locations and list of ref types (either REF or SHAM, defined in CitationFinder)
"""
refs = []
locations = []
types = []
failed_refs = []
# failed_non_refs = []
failed_shams = []
assert lang == 'he'
# todo: support english
st = strip_nikkud(st)
all_refs = self._citationFinder.get_potential_refs(st, lang)
for item, location, type in all_refs:
if type == CitationFinder.REF_INT:
try:
refs += [self._tr.resolve(item.index_node.full_title(), item.sections)]
locations += [location]
types += [type]
#refs += [(self._tr.resolve(item.index_node.full_title(), item.sections), 'ref')]
except (IbidRefException, IbidKeyNotFoundException) as e:
failed_refs += [item.normal()]
elif type == CitationFinder.NON_REF_INT or type == CitationFinder.IGNORE_INT:
# failed_non_refs += [item.group()]
self._tr.ignore_book_name_keys()
elif type == CitationFinder.SHAM_INT:
try:
if isinstance(item, unicode):
refs += [self._tr.resolve(None, match_str=item)]
locations += [location]
types += [type]
#refs += [(self._tr.resolve(None, match_str=item), 'sham')]
else:
refs += [self._tr.resolve(item[0], sections=item[1])]
locations += [location]
types += [type]
#refs += [(self._tr.resolve(item[0], sections=item[1]), 'sham')]
except (IbidRefException, IbidKeyNotFoundException, InputError) as e:
failed_shams += [item]
return refs, locations, types # , failed_refs, failed_non_refs, failed_shams
def body(self):
self.load_toc()
self.click_toc_category("Midrash")
self.click_toc_text("Ein Yaakov")
self.click_source_title()
self.click_masechet_and_chapter('2', '3')
section = self.get_section_txt('1')
assert 'רבי זירא הוה' in strip_nikkud(section)
self.load_toc()
self.click_toc_category("Midrash").click_toc_text("Midrash Mishlei")
self.click_source_title()
self.click_chapter('4')
section = self.get_section_txt('1')
assert 'מכל משמר נצור ליבך' in section
def create_section(oref, dicta_text, dicta_vtitle):
with_nikkud, without_nukkud = dicta_text.split(), strip_nikkud(
dicta_text).split()
sefaria_text = prepare_sefaria_text(oref)
dh_match = match_text(without_nukkud, sefaria_text)
matches = dh_match['matches']
segments = oref.all_segment_refs()
assert len(segments) == len(matches)
for segment, match in zip(segments, matches):
tc = segment.text('he', dicta_vtitle)
new_segment_text = u' '.join(with_nikkud[match[0]:match[1] + 1])
if not new_segment_text:
new_segment_text = segment.text('he', davidson_vtitle).text
tc.text = new_segment_text
tc.save()
def count_terms(query={}, lang=None):
#todo: move to object model. Maybe. What's this doing?
"""
Counts all terms in texts matching query, lang
Saves reults to terms collection in db.
"""
terms = {}
bavli_names = db.index.find(query).distinct("title")
query = {"title": {"$in": bavli_names}}
refs = counts.generate_refs_list(
query) #library.ref_list() needs query argument
lookup_lang = "he" if lang == "ar" else lang
for ref in refs:
print(ref)
#text = texts.get_text(ref, commentary=False)
text = TextFamily(Ref(ref), commentary=False).contents()
for i, line in enumerate(text.get(lookup_lang, [])):
# strip punctuation
for c in string.punctuation:
line = line.replace(c, "")
these_terms = line.split(" ")
for term in these_terms:
line_ref = "%s:%d" % (ref, i + 1)
term = hebrew.strip_nikkud(term)
if term in terms:
terms[term]["occurrences"] += 1
terms[term]["refs"].add(line_ref)
else:
terms[term] = {
"term": term,
"occurrences": 1,
"language": lang,
"refs": set([line_ref])
}
for term in terms:
print(term)
# only include up to 20 random ref samples
sample_size = len(
terms[term]["refs"]) if len(terms[term]["refs"]) < 20 else 20
terms[term]["refs"] = list(sample(terms[term]["refs"], sample_size))
db.terms.save(terms[term])
def count_terms(query={}, lang=None):
#todo: move to object model. Maybe. What's this doing?
"""
Counts all terms in texts matching query, lang
Saves reults to terms collection in db.
"""
terms = {}
bavli_names = db.index.find(query).distinct("title")
query = {"title": {"$in": bavli_names}}
refs = counts.generate_refs_list(query) #library.ref_list() needs query argument
lookup_lang = "he" if lang == "ar" else lang
for ref in refs:
print ref
#text = texts.get_text(ref, commentary=False)
text = TextFamily(Ref(ref), commentary=False).contents()
for i, line in enumerate(text.get(lookup_lang, [])):
# strip punctuation
for c in string.punctuation:
line = line.replace(c,"")
these_terms = line.split(" ")
for term in these_terms:
line_ref = "%s:%d" % (ref, i+1)
term = hebrew.strip_nikkud(term)
if term in terms:
terms[term]["occurrences"] += 1
terms[term]["refs"].add(line_ref)
else:
terms[term] = {
"term": term,
"occurrences": 1,
"language": lang,
"refs": set([line_ref])
}
for term in terms:
print term
# only include up to 20 random ref samples
sample_size = len(terms[term]["refs"]) if len(terms[term]["refs"]) < 20 else 20
terms[term]["refs"] = list(sample(terms[term]["refs"], sample_size))
db.terms.save(terms[term])
def find_all_shams_in_st(self, st, lang = 'he'):
'''
:param st: source text
:param lang:
:return: a list of tuples (Refs objects that originally were Shams, location)
'''
from sefaria.utils.hebrew import strip_nikkud
st = strip_nikkud(st)
sham_refs = []
reg = u'(\(|\([^)]* )שם(\)| [^(]*\))' # finds shams in parenthesis without רבשם
for sham in re.finditer(reg, st):
matched = sham.group()
if len(re.split('\s+', matched)) > 6: # todo: find stitistics of the cutoff size of a ref-citation 6 is a guess
continue
try:
sham_refs += [(self.parse_sham(matched),sham.span())]
except IbidKeyNotFoundException:
pass
except IbidRefException:
pass # maybe want to call ignore her?
return sham_refs
def find_all_shams_in_st(self, st, lang = 'he'):
'''
:param st: source text
:param lang:
:return: a list of tuples (Refs objects that originally were Shams, location)
'''
from sefaria.utils.hebrew import strip_nikkud
st = strip_nikkud(st)
sham_refs = []
reg = u'(\(|\([^)]* )שם(\)| [^(]*\))' # finds shams in parenthesis without רבשם
for sham in re.finditer(reg, st):
matched = sham.group()
if len(re.split('\s+', matched)) > 6: # todo: find stitistics of the cutoff size of a ref-citation 6 is a guess
continue
try:
sham_refs += [(self.parse_sham(matched),sham.span())]
except IbidKeyNotFoundException:
pass
except IbidRefException:
pass # maybe want to call ignore her?
return sham_refs
def run_shaminator(titles=None, with_real_refs=False, SEG_DIST = 5, create_ref_dict = True):
base_url = u"https://www.sefaria.org/"
title_list = []
cats = ["Midrash", "Halakha", "Philosophy"]
collective_titles = ["Rashi", "Kessef Mishneh"]
for cat in cats:
title_list += library.get_indexes_in_category(cat)
for cTitle in collective_titles:
title_list += library.get_indices_by_collective_title(cTitle)
title_list = titles
for ititle, title in enumerate(title_list):
print u"-"*50
print title, ititle+1, '/', len(title_list)
print u"-"*50
html = u"""
<!DOCTYPE html>
<html>
<head>
<link rel='stylesheet' type='text/css' href='styles.css'>
<meta charset='utf-8'>
</head>
<body>
<table>
<tr><td>Row Id</td><td>Book Ref</td><td>Ref Found</td><td>Sham Found</td><td>Sham Text</td></tr>
"""
index = library.get_index(title)
inst = IndexIbidFinder(index)
if create_ref_dict:
try:
ref_dict = inst.find_in_index()
# ref_dict - OrderedDict. keys: segments. values: dict {'refs': [Refs obj found in this seg], 'locations': [], 'types': []}
except AssertionError:
print "Skipping {}".format(title)
continue # problem with Ein Ayah
last_index_ref_seen = {}
row_num = 1
char_padding = 20
double_tanakh_books = {"I Samuel": "Samuel", "II Samuel": "Samuel", "I Kings": "Kings", "II Kings": "Kings",
"I Chronicles": "Chronicles", "II Chronicles": "Chronicles"}
for k, v in ref_dict.items():
curr_ref = Ref(k)
for i, (r, l, t) in enumerate(izip(v['refs'], v['locations'], v['types'])):
sham_ref_key = r.index.title if r.index.title not in double_tanakh_books else double_tanakh_books[
r.index.title]
if t == CitationFinder.SHAM_INT and last_index_ref_seen[sham_ref_key] is not None:
last_ref_with_citation, last_location_with_citation, last_ref_seen = last_index_ref_seen[sham_ref_key]
else: # if t == CitationFinder.REF_INT:
last_index_ref_seen[sham_ref_key] = (curr_ref, l, r)
if not with_real_refs:
continue
dist = curr_ref.distance(last_ref_with_citation)
last_ref_with_citation = curr_ref
last_location_with_citation = l
last_ref_seen = r
r = u"N/A"
# dist = curr_ref.distance(last_ref_with_citation)
print dist
if dist == 0:
text = strip_nikkud(curr_ref.text('he').text)
start_ind = 0 if last_location_with_citation[0] - char_padding < 0 else last_location_with_citation[
0] - char_padding
end_ind = l[1] + char_padding
before = text[start_ind:last_location_with_citation[0]]
real_ref = text[last_location_with_citation[0]:last_location_with_citation[1]]
middle = text[last_location_with_citation[1]:l[0]] if last_location_with_citation[1] <= l[0] else u""
sham_ref = text[l[0]:l[1]] if t == CitationFinder.SHAM_INT else u""
after = text[l[1]:end_ind]
text = u"{}<span class='r'>{}</span>{}<span class='s'>{}</span>{}".format(before, real_ref, middle,
sham_ref, after)
else:
start_text = strip_nikkud(last_ref_with_citation.text('he').text)
# start_text = strip_nikkud(start_text)[last_location_with_citation[0]:]
end_text = strip_nikkud(curr_ref.text('he').text)
# end_text = strip_nikkud(end_text)[:l[1]+1]
if dist > SEG_DIST:
continue
elif dist > 1 and dist <= SEG_DIST:
print u"{} {} {}".format(curr_ref, last_ref_with_citation.next_segment_ref(),
curr_ref.prev_segment_ref())
mid_text = last_ref_with_citation.next_segment_ref().to(curr_ref.prev_segment_ref()).text(
'he').text
while isinstance(mid_text, list):
mid_text = reduce(lambda a, b: a + b, mid_text)
else:
mid_text = u""
start_ind = 0 if last_location_with_citation[0] - char_padding < 0 else last_location_with_citation[
0] - char_padding
end_ind = l[1] + char_padding
start_before = start_text[start_ind:last_location_with_citation[0]]
start_real_ref = start_text[last_location_with_citation[0]:last_location_with_citation[1]]
start_after = start_text[last_location_with_citation[1]:]
end_before = end_text[:l[0]]
end_sham_ref = end_text[l[0]:l[1]]
end_after = end_text[l[1]:end_ind]
text = u"{}<span class='r'>{}</span>{} {} {}<span class='s'>{}</span>{}".format(start_before,
start_real_ref,
start_after,
mid_text,
end_before,
end_sham_ref,
end_after)
text = bleach.clean(text, strip=True, tags=[u'span'], attributes=[u'class'])
# surround all non interesting parens with spans
text = re.sub(ur"(?<!>)(\([^)]+\))(?!<)", ur"<span class='p'>\1</span>", text)
rowclass = u"realrefrow" if t == CitationFinder.REF_INT else u"shamrefrow"
row = u"<tr class='{}' ><td>{}</td><td><a href='{}' target='_blank'>{}</a></td><td>{}</td><td>{}</td><td class='he'>{}</td></tr>"\
.format(rowclass, row_num, base_url + curr_ref.url(), k, last_ref_seen, r, text)
html += row
row_num += 1
html += u"""
</table>
</body>
</html>
"""
with codecs.open('ibid_output/ibid_{}.html'.format(title), 'wb', encoding='utf8') as f:
f.write(html)
import sefaria.tracker as tracker
patterns = [u"כתרגומו", u"ותרגומו", u"תרגומו", u"וזהו שתרגם אונקלוס", u"אונקלוס", u"לכך מתרגם", u"מתרגם"]
books = ["Genesis", "Exodus", "Leviticus", "Numbers", "Deuteronomy"]
total = 0
for book in books:
rashi_book = "Rashi on " + book
onkelos_book = "Onkelos " + book
i = library.get_index(rashi_book)
assert isinstance(i, CommentaryIndex)
all_rashis = i.all_segment_refs()
# Loop through all of the Rashis
for rashi_ref in all_rashis:
rashi = strip_nikkud(TextChunk(rashi_ref, "he", "On Your Way").text)
# If it matches the pattern
for pat in patterns:
if pat in rashi:
onkelos_ref = Ref(rashi_ref.section_ref().normal().replace(rashi_book, onkelos_book))
d = {
"refs": [rashi_ref.normal(), onkelos_ref.normal()],
"type": "reference",
"auto": True,
"generated_by": "Rashi - Onkelos Linker",
}
tracker.add(28, Link, d)
print u"{}\t{}\t{}".format(rashi_ref.normal(), pat, rashi.strip())
total += 1
break
def test_strip_nikkud(self):
assert h.strip_nikkud('הַדְּבָרִים אֲשֶׁר') == 'הדברים אשר'
assert h.strip_nikkud(
"הַמּוֹצִיא בְמִסְפָּר צְבָאָם לְכֻלָּם בְּשֵׁם יִקְרָא"
) == "המוציא במספר צבאם לכלם בשם יקרא"
def run_shaminator(titles=None, with_real_refs=False):
base_url = u"https://www.sefaria.org/"
title_list = []
cats = ["Midrash", "Halakha", "Philosophy"]
collective_titles = ["Rashi", "Kessef Mishneh"]
for cat in cats:
title_list += library.get_indexes_in_category(cat)
for cTitle in collective_titles:
title_list += library.get_indices_by_collective_title(cTitle)
title_list = titles
for ititle, title in enumerate(title_list):
print u"-"*50
print title, ititle+1, '/', len(title_list)
print u"-"*50
html = u"""
<!DOCTYPE html>
<html>
<head>
<link rel='stylesheet' type='text/css' href='styles.css'>
<meta charset='utf-8'>
</head>
<body>
<table>
<tr><td>Row Id</td><td>Book Ref</td><td>Ref Found</td><td>Sham Found</td><td>Sham Text</td></tr>
"""
index = library.get_index(title)
inst = IndexIbidFinder(index)
try:
ref_dict = inst.find_in_index()
except AssertionError:
print "Skipping {}".format(title)
continue # problem with Ein Ayah
last_index_ref_seen = {}
row_num = 1
char_padding = 20
double_tanakh_books = {"I Samuel": "Samuel", "II Samuel": "Samuel", "I Kings": "Kings", "II Kings": "Kings",
"I Chronicles": "Chronicles", "II Chronicles": "Chronicles"}
for k, v in ref_dict.items():
curr_ref = Ref(k)
for r, l, t in izip(v['refs'], v['locations'], v['types']):
sham_ref_key = r.index.title if r.index.title not in double_tanakh_books else double_tanakh_books[
r.index.title]
if t == CitationFinder.SHAM_INT and last_index_ref_seen[sham_ref_key] is not None:
last_ref_with_citation, last_location_with_citation, last_ref_seen = last_index_ref_seen[sham_ref_key]
else: # if t == CitationFinder.REF_INT:
last_index_ref_seen[sham_ref_key] = (curr_ref, l, r)
if not with_real_refs:
continue
last_ref_with_citation = curr_ref
last_location_with_citation = l
last_ref_seen = r
r = u"N/A"
dist = curr_ref.distance(last_ref_with_citation)
if dist == 0:
text = strip_nikkud(curr_ref.text('he').text)
start_ind = 0 if last_location_with_citation[0] - char_padding < 0 else last_location_with_citation[
0] - char_padding
end_ind = l[1] + char_padding
before = text[start_ind:last_location_with_citation[0]]
real_ref = text[last_location_with_citation[0]:last_location_with_citation[1]]
middle = text[last_location_with_citation[1]:l[0]] if last_location_with_citation[1] <= l[0] else u""
sham_ref = text[l[0]:l[1]] if t == CitationFinder.SHAM_INT else u""
after = text[l[1]:end_ind]
text = u"{}<span class='r'>{}</span>{}<span class='s'>{}</span>{}".format(before, real_ref, middle,
sham_ref, after)
else:
start_text = strip_nikkud(last_ref_with_citation.text('he').text)
# start_text = strip_nikkud(start_text)[last_location_with_citation[0]:]
end_text = strip_nikkud(curr_ref.text('he').text)
# end_text = strip_nikkud(end_text)[:l[1]+1]
if dist > 1:
print u"{} {} {}".format(curr_ref, last_ref_with_citation.next_segment_ref(),
curr_ref.prev_segment_ref())
mid_text = last_ref_with_citation.next_segment_ref().to(curr_ref.prev_segment_ref()).text(
'he').text
while isinstance(mid_text, list):
mid_text = reduce(lambda a, b: a + b, mid_text)
else:
mid_text = u""
start_ind = 0 if last_location_with_citation[0] - char_padding < 0 else last_location_with_citation[
0] - char_padding
end_ind = l[1] + char_padding
start_before = start_text[start_ind:last_location_with_citation[0]]
start_real_ref = start_text[last_location_with_citation[0]:last_location_with_citation[1]]
start_after = start_text[last_location_with_citation[1]:]
end_before = end_text[:l[0]]
end_sham_ref = end_text[l[0]:l[1]]
end_after = end_text[l[1]:end_ind]
text = u"{}<span class='r'>{}</span>{} {} {}<span class='s'>{}</span>{}".format(start_before,
start_real_ref,
start_after,
mid_text,
end_before,
end_sham_ref,
end_after)
text = bleach.clean(text, strip=True, tags=[u'span'], attributes=[u'class'])
# surround all non interesting parens with spans
text = re.sub(ur"(?<!>)(\([^)]+\))(?!<)", ur"<span class='p'>\1</span>", text)
rowclass = u"realrefrow" if t == CitationFinder.REF_INT else u"shamrefrow"
row = u"<tr class='{}' ><td>{}</td><td><a href='{}' target='_blank'>{}</a></td><td>{}</td><td>{}</td><td class='he'>{}</td></tr>"\
.format(rowclass, row_num, base_url + curr_ref.url(), k, last_ref_seen, r, text)
html += row
row_num += 1
html += u"""
</table>
</body>
</html>
"""
with codecs.open('ibid_output/ibid_{}.html'.format(title), 'wb',encoding='utf8') as f:
f.write(html)
for sheet in sheet_list:
for tag in sheet.get("tags", []):
original_tag_counter[tag] += 1
sorted_tags = sorted(original_tag_counter,
key=original_tag_counter.get,
reverse=True)
sorted_en_tags = [t for t in sorted_tags if not is_hebrew(t)]
sorted_he_tags = [t for t in sorted_tags if is_hebrew(t)]
translated_hebrew_tags = defaultdict(list)
for en_tag in sorted_en_tags:
translation = translate_client.translate(en_tag,
target_language='iw',
source_language='en')
he_tag = strip_nikkud(translation['translatedText'])
if en_tag == he_tag:
print("Couldn't translate {}".format(en_tag))
untranslated_en_tags += [en_tag]
continue
print("{}:{}".format(he_tag, en_tag))
translated_hebrew_tags[he_tag] += [en_tag]
overall_counts = {
he_tag: sum([
original_tag_counter[en_tag]
for en_tag in translated_hebrew_tags[he_tag]
])
for he_tag in translated_hebrew_tags
}
ordered_translated_he_terms = sorted(overall_counts,
u"כתרגומו", u"ותרגומו", u"תרגומו", u"וזהו שתרגם אונקלוס", u"אונקלוס",
u"לכך מתרגם", u"מתרגם"
]
books = ["Genesis", "Exodus", "Leviticus", "Numbers", "Deuteronomy"]
total = 0
for book in books:
rashi_book = "Rashi on " + book
onkelos_book = "Onkelos " + book
i = library.get_index(rashi_book)
assert isinstance(i, CommentaryIndex)
all_rashis = i.all_segment_refs()
# Loop through all of the Rashis
for rashi_ref in all_rashis:
rashi = strip_nikkud(TextChunk(rashi_ref, "he", "On Your Way").text)
# If it matches the pattern
for pat in patterns:
if pat in rashi:
onkelos_ref = Ref(rashi_ref.section_ref().normal().replace(
rashi_book, onkelos_book))
d = {
"refs": [rashi_ref.normal(),
onkelos_ref.normal()],
"type": "reference",
"auto": True,
"generated_by": "Rashi - Onkelos Linker"
}
tracker.add(28, Link, d)
print u"{}\t{}\t{}".format(rashi_ref.normal(), pat,
def text_strip_nikkud(self):
assert h.strip_nikkud(u'הַדְּבָרִים אֲשֶׁר') == u'הדברים אשר'
def test_strip_nikkud(self):
assert h.strip_nikkud(u'הַדְּבָרִים אֲשֶׁר') == u'הדברים אשר'
assert h.strip_nikkud(u"הַמּוֹצִיא בְמִסְפָּר צְבָאָם לְכֻלָּם בְּשֵׁם יִקְרָא") == u"המוציא במספר צבאם לכלם בשם יקרא"
def text_strip_nikkud(self):
assert h.strip_nikkud(u'הַדְּבָרִים אֲשֶׁר') == u'הדברים אשר'
