def load_comparison(left_bill, left_version, right_bill, right_version, timelimit=10, force=False):
from billtext import load_bill_text, compare_xml_text, get_current_version
import lxml
left_bill = Bill.objects.get(id = left_bill)
right_bill = Bill.objects.get(id = right_bill)
if left_version == "": left_version = get_current_version(left_bill)
if right_version == "": right_version = get_current_version(right_bill)
btc = None
try:
btc = BillTextComparison.objects.get(
bill1 = left_bill,
ver1 = left_version,
bill2 = right_bill,
ver2 = right_version)
btc.decompress()
if not force: return btc.data
except BillTextComparison.DoesNotExist:
pass
# Try with the bills swapped.
try:
btc2 = BillTextComparison.objects.get(
bill2 = left_bill,
ver2 = left_version,
bill1 = right_bill,
ver1 = right_version)
btc2.decompress()
data = btc2.data
return {
"left_meta": data["right_meta"],
"right_meta": data["left_meta"],
"left_text": data["right_text"],
"right_text": data["left_text"],
}
except BillTextComparison.DoesNotExist:
pass
left = load_bill_text(left_bill, left_version, mods_only=True)
right = load_bill_text(right_bill, right_version, mods_only=True)
doc1 = lxml.etree.parse(left["basename"] + ".html")
doc2 = lxml.etree.parse(right["basename"] + ".html")
compare_xml_text(doc1, doc2, timelimit=timelimit) # revises DOMs in-place
# dates aren't JSON serializable
left["docdate"] = left["docdate"].strftime("%x")
right["docdate"] = right["docdate"].strftime("%x")
ret = {
"left_meta": left,
"right_meta": right,
"left_text": lxml.etree.tostring(doc1),
"right_text": lxml.etree.tostring(doc2),
}
if not btc:
btc = BillTextComparison(
bill1 = left_bill,
ver1 = left_version,
bill2 = right_bill,
ver2 = right_version,
data = dict(ret)) # clone before compress()
else:
btc.data = dict(ret) # clone before compress()
btc.compress()
btc.save()
return ret
def load_comparison(left_bill, left_version, right_bill, right_version, timelimit=10, use_cache=True, force_update=False):
from .billtext import load_bill_text, get_current_version
from xml_diff import compare
import lxml
left_bill = Bill.objects.get(id = left_bill)
right_bill = Bill.objects.get(id = right_bill)
if left_version == "": left_version = get_current_version(left_bill)
if right_version == "": right_version = get_current_version(right_bill)
if use_cache:
# Load from cache.
try:
btc = BillTextComparison.objects.get(
bill1 = left_bill,
ver1 = left_version,
bill2 = right_bill,
ver2 = right_version)
btc.decompress()
return btc.data
except BillTextComparison.DoesNotExist:
pass
# Load from cache - Try with the bills swapped.
try:
btc2 = BillTextComparison.objects.get(
bill2 = left_bill,
ver2 = left_version,
bill1 = right_bill,
ver1 = right_version)
btc2.decompress()
data = btc2.data
# un-swap
return {
"left_meta": data["right_meta"],
"right_meta": data["left_meta"],
"left_text": data["right_text"],
"right_text": data["left_text"],
}
except BillTextComparison.DoesNotExist:
pass
# Load bill text metadata.
left = load_bill_text(left_bill, left_version, mods_only=True)
right = load_bill_text(right_bill, right_version, mods_only=True)
# Load XML DOMs for each document and perform the comparison.
def load_bill_text_xml(docinfo):
# If XML text is available, use it, but pre-render it
# into HTML. Otherwise use the legacy HTML that we
# scraped from THOMAS.
if "xml_file" in docinfo:
import congressxml
return congressxml.convert_xml(docinfo["xml_file"])
elif "html_file" in docinfo:
return lxml.etree.parse(docinfo["html_file"])
else:
raise IOError("Bill text is not available for one of the bills.")
doc1 = load_bill_text_xml(left)
doc2 = load_bill_text_xml(right)
def make_tag_func(ins_del):
import lxml.etree
elem = lxml.etree.Element("comparison-change")
return elem
def differ(text1, text2):
# ensure we use the C++ Google DMP and can specify the time limit
import diff_match_patch
for x in diff_match_patch.diff(text1, text2, timelimit=timelimit):
yield x
compare(doc1.getroot(), doc2.getroot(), make_tag_func=make_tag_func, differ=differ)
# Prepare JSON response data.
# dates aren't JSON serializable
left["docdate"] = left["docdate"].strftime("%x")
right["docdate"] = right["docdate"].strftime("%x")
ret = {
"left_meta": left,
"right_meta": right,
"left_text": lxml.etree.tostring(doc1, encoding=str),
"right_text": lxml.etree.tostring(doc2, encoding=str),
}
if use_cache or force_update:
# For force_update, or race conditions, delete any existing record.
fltr = { "bill1": left_bill,
"ver1": left_version,
"bill2": right_bill,
"ver2": right_version }
BillTextComparison.objects.filter(**fltr).delete()
# Cache in database so we don't have to re-do the comparison
# computation again.
btc = BillTextComparison(
data = dict(ret), # clone before compress()
**fltr)
btc.compress()
btc.save()
# Return JSON comparison data.
return ret
def load_comparison(left_bill,
left_version,
right_bill,
right_version,
timelimit=10):
from billtext import load_bill_text, get_current_version
from xml_diff import compare
import lxml
left_bill = Bill.objects.get(id=left_bill)
right_bill = Bill.objects.get(id=right_bill)
if left_version == "": left_version = get_current_version(left_bill)
if right_version == "": right_version = get_current_version(right_bill)
use_cache = True
if use_cache:
# Load from cache.
try:
btc = BillTextComparison.objects.get(bill1=left_bill,
ver1=left_version,
bill2=right_bill,
ver2=right_version)
btc.decompress()
return btc.data
except BillTextComparison.DoesNotExist:
pass
# Load from cache - Try with the bills swapped.
try:
btc2 = BillTextComparison.objects.get(bill2=left_bill,
ver2=left_version,
bill1=right_bill,
ver1=right_version)
btc2.decompress()
data = btc2.data
# un-swap
return {
"left_meta": data["right_meta"],
"right_meta": data["left_meta"],
"left_text": data["right_text"],
"right_text": data["left_text"],
}
except BillTextComparison.DoesNotExist:
pass
# Load bill text metadata.
left = load_bill_text(left_bill, left_version, mods_only=True)
right = load_bill_text(right_bill, right_version, mods_only=True)
# Load XML DOMs for each document and perform the comparison.
def load_bill_text_xml(docinfo):
# If XML text is available, use it, but pre-render it
# into HTML. Otherwise use the legacy HTML that we
# scraped from THOMAS.
if "xml_file" in docinfo:
import congressxml
return congressxml.convert_xml(docinfo["xml_file"])
elif "html_file" in docinfo:
return lxml.etree.parse(docinfo["html_file"])
else:
raise IOError("Bill text is not available for one of the bills.")
doc1 = load_bill_text_xml(left)
doc2 = load_bill_text_xml(right)
def make_tag_func(ins_del):
import lxml.etree
elem = lxml.etree.Element("comparison-change")
return elem
compare(doc1.getroot(), doc2.getroot(), make_tag_func=make_tag_func)
# Prepare JSON response data.
# dates aren't JSON serializable
left["docdate"] = left["docdate"].strftime("%x")
right["docdate"] = right["docdate"].strftime("%x")
ret = {
"left_meta": left,
"right_meta": right,
"left_text": lxml.etree.tostring(doc1),
"right_text": lxml.etree.tostring(doc2),
}
if use_cache:
# Cache in database so we don't have to re-do the comparison
# computation again.
btc = BillTextComparison(bill1=left_bill,
ver1=left_version,
bill2=right_bill,
ver2=right_version,
data=dict(ret)) # clone before compress()
btc.compress()
btc.save()
# Return JSON comparison data.
return ret
def load_comparison(left_bill,
left_version,
right_bill,
right_version,
timelimit=10,
force=False):
from billtext import load_bill_text, compare_xml_text, get_current_version
import lxml
left_bill = Bill.objects.get(id=left_bill)
right_bill = Bill.objects.get(id=right_bill)
if left_version == "": left_version = get_current_version(left_bill)
if right_version == "": right_version = get_current_version(right_bill)
btc = None
try:
btc = BillTextComparison.objects.get(bill1=left_bill,
ver1=left_version,
bill2=right_bill,
ver2=right_version)
btc.decompress()
if not force: return btc.data
except BillTextComparison.DoesNotExist:
pass
# Try with the bills swapped.
try:
btc2 = BillTextComparison.objects.get(bill2=left_bill,
ver2=left_version,
bill1=right_bill,
ver1=right_version)
btc2.decompress()
data = btc2.data
return {
"left_meta": data["right_meta"],
"right_meta": data["left_meta"],
"left_text": data["right_text"],
"right_text": data["left_text"],
}
except BillTextComparison.DoesNotExist:
pass
left = load_bill_text(left_bill, left_version, mods_only=True)
right = load_bill_text(right_bill, right_version, mods_only=True)
try:
doc1 = lxml.etree.parse(left["html_file"])
doc2 = lxml.etree.parse(right["html_file"])
except KeyError:
raise IOError(
"The HTML bill text format is not available for one of the bills.")
compare_xml_text(doc1, doc2, timelimit=timelimit) # revises DOMs in-place
# dates aren't JSON serializable
left["docdate"] = left["docdate"].strftime("%x")
right["docdate"] = right["docdate"].strftime("%x")
ret = {
"left_meta": left,
"right_meta": right,
"left_text": lxml.etree.tostring(doc1),
"right_text": lxml.etree.tostring(doc2),
}
if not btc:
btc = BillTextComparison(bill1=left_bill,
ver1=left_version,
bill2=right_bill,
ver2=right_version,
data=dict(ret)) # clone before compress()
else:
btc.data = dict(ret) # clone before compress()
btc.compress()
btc.save()
return ret
def load_comparison(left_bill, left_version, right_bill, right_version, timelimit=10, use_cache=True, force_update=False):
from billtext import load_bill_text, get_current_version
from xml_diff import compare
import lxml
left_bill = Bill.objects.get(id = left_bill)
right_bill = Bill.objects.get(id = right_bill)
if left_version == "": left_version = get_current_version(left_bill)
if right_version == "": right_version = get_current_version(right_bill)
if use_cache:
# Load from cache.
try:
btc = BillTextComparison.objects.get(
bill1 = left_bill,
ver1 = left_version,
bill2 = right_bill,
ver2 = right_version)
btc.decompress()
return btc.data
except BillTextComparison.DoesNotExist:
pass
# Load from cache - Try with the bills swapped.
try:
btc2 = BillTextComparison.objects.get(
bill2 = left_bill,
ver2 = left_version,
bill1 = right_bill,
ver1 = right_version)
btc2.decompress()
data = btc2.data
# un-swap
return {
"left_meta": data["right_meta"],
"right_meta": data["left_meta"],
"left_text": data["right_text"],
"right_text": data["left_text"],
}
except BillTextComparison.DoesNotExist:
pass
# Load bill text metadata.
left = load_bill_text(left_bill, left_version, mods_only=True)
right = load_bill_text(right_bill, right_version, mods_only=True)
# Load XML DOMs for each document and perform the comparison.
def load_bill_text_xml(docinfo):
# If XML text is available, use it, but pre-render it
# into HTML. Otherwise use the legacy HTML that we
# scraped from THOMAS.
if "xml_file" in docinfo:
import congressxml
return congressxml.convert_xml(docinfo["xml_file"])
elif "html_file" in docinfo:
return lxml.etree.parse(docinfo["html_file"])
else:
raise IOError("Bill text is not available for one of the bills.")
doc1 = load_bill_text_xml(left)
doc2 = load_bill_text_xml(right)
def make_tag_func(ins_del):
import lxml.etree
elem = lxml.etree.Element("comparison-change")
return elem
def differ(text1, text2):
# ensure we use the C++ Google DMP and can specify the time limit
import diff_match_patch
for x in diff_match_patch.diff_unicode(text1, text2, timelimit=timelimit):
yield x
compare(doc1.getroot(), doc2.getroot(), make_tag_func=make_tag_func, differ=differ)
# Prepare JSON response data.
# dates aren't JSON serializable
left["docdate"] = left["docdate"].strftime("%x")
right["docdate"] = right["docdate"].strftime("%x")
ret = {
"left_meta": left,
"right_meta": right,
"left_text": lxml.etree.tostring(doc1),
"right_text": lxml.etree.tostring(doc2),
}
if use_cache or force_update:
# For force_update, or race conditions, delete any existing record.
fltr = { "bill1": left_bill,
"ver1": left_version,
"bill2": right_bill,
"ver2": right_version }
BillTextComparison.objects.filter(**fltr).delete()
# Cache in database so we don't have to re-do the comparison
# computation again.
btc = BillTextComparison(
data = dict(ret), # clone before compress()
**fltr)
btc.compress()
btc.save()
# Return JSON comparison data.
return ret
