def run(options):
year_range = inspector.year_range(options, archive)
pages = options.get('pages', ALL_PAGES)
# Pull the audit reports. Pages are 0-indexed.
for page in range(0, int(pages) - 1):
doc = BeautifulSoup(utils.download(AUDIT_REPORTS_URL.format(page=page)))
results = doc.select("span.field-content")
if not results:
# No more results, we must have hit the last page
break
for result in results:
report = report_from(result, year_range, report_type='audit')
if report:
inspector.save_report(report)
# Grab the other reports
for report_type, url in OTHER_REPORT_URLS.items():
doc = BeautifulSoup(utils.download(url))
results = doc.select(".views-field")
if not results:
results = doc.select(".views-row")
for result in results:
report = report_from(result, year_range, report_type)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
for year in year_range:
if year < 2006: # The oldest year for audit reports
continue
url = AUDIT_REPORTS_URL.format(year=year)
doc = BeautifulSoup(utils.download(url))
results = doc.select("div#content li")
for result in results:
report = audit_report_from(result, url, year, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.select("div#content li")
for result in results:
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the Peer Review
doc = BeautifulSoup(utils.download(PEER_REVIEWS_URL))
result = doc.find("div", id='content').find("a", text=True)
report = peer_review_from(result, year_range)
inspector.save_report(report)
def urls_for_topics(self, topics):
for topic in topics:
# Topic might be a tuple for ADDITIONAL_TOPICS (not ones from command
# line).
self.report_type = None
if isinstance(topic, tuple):
topic, report_type = topic
self.report_type = report_type
last_page = False
url = TOPIC_TO_URL[topic]
page = BeautifulSoup(utils.download(url))
page_started = self.is_first_page(page)
if page_started:
yield url
for link in page.select('li.pager-item a'):
next_url = urljoin(url, link['href'])
next_page = BeautifulSoup(utils.download(next_url))
if not page_started:
page_started = self.is_first_page(next_page)
if page_started:
yield next_url
last_page = self.is_last_page(next_page)
if last_page:
break
if last_page:
continue
self.report_type = None # Clear this out afterwards
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
for year in year_range:
url = AUDITS_REPORTS_URL.format(str(year)[2:4])
doc = BeautifulSoup(utils.download(url))
results = doc.select("tr")
if not results:
raise inspector.NoReportsFoundError("NASA (%d)" % year)
for index, result in enumerate(results):
if not index or not result.text.strip():
# Skip the header row and any empty rows
continue
report = audit_report_from(result, url, year_range)
if report:
inspector.save_report(report)
# Pull the other reports
doc = BeautifulSoup(utils.download(OTHER_REPORT_URL))
results = doc.select("#subContainer ul li")
if not results:
raise inspector.NoReportsFoundError("NASA (other)")
for result in results:
report = other_report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# # Pull the RSS feed
doc = BeautifulSoup(utils.download(RSS_URL))
results = doc.select("item")
for result in results:
report = rss_report_from(result, year_range)
if report:
inspector.save_report(report)
# # Pull the recent audit reports.
doc = BeautifulSoup(utils.download(RECENT_AUDITS_URL))
results = doc.select("div.block > a")
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the archive audit reports
doc = BeautifulSoup(utils.download(AUDIT_ARCHIVE_URL))
results = doc.select("div.block a")
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the other reports
doc = BeautifulSoup(utils.download(OTHER_REPORTS_URl))
results = doc.select("div.block > a")
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
for year in year_range:
if year < 2005: # This is the earliest audits go back
continue
url = AUDIT_REPORTS_URL.format(year=year)
doc = BeautifulSoup(utils.download(url))
results = doc.select("div.content")
if not results:
raise inspector.NoReportsFoundError("Tennessee Valley Authority (%d)" % year)
for result in results:
report = audit_report_from(result, url, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.select("report")
if not results:
raise inspector.NoReportsFoundError("Tennessee Valley Authority (semiannual reports)")
for result in results:
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the reports with pagination
for report_type, report_url_format in PAGINATED_REPORT_FORMATS.items():
for page in range(0, 999):
url = report_url_format.format(page=page)
doc = BeautifulSoup(utils.download(url))
results = doc.select("li.views-row")
if not results:
if page == 0:
raise inspector.NoReportsFoundError("USAID (%s)" % report_type)
else:
break
for result in results:
report = report_from(result, url, report_type, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports (no pagination)
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.select("li.views-row")
if not results:
raise inspector.NoReportsFoundError("USAID (semiannual reports)")
for result in results:
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the general reports
doc = BeautifulSoup(utils.download(REPORTS_URL))
results = doc.select("div#mainContent li.mainContenttext a")
for result in results:
report = report_from(result, REPORTS_URL, year_range)
if report:
inspector.save_report(report)
# Pull the archive reports
doc = BeautifulSoup(utils.download(REPORT_ARCHIVE_URL))
results = doc.select("div#mainContent li.mainContenttext a") + doc.select("div#mainContent span.mainContenttext a")
for result in results:
if not result.text:
continue
report = report_from(result, REPORT_ARCHIVE_URL, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.select("div#mainContent li.mainContenttext a")
for result in results:
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
for year in year_range:
if year < 2002: # The oldest page for audit reports
continue
doc = BeautifulSoup(utils.download(AUDIT_REPORTS_URL.format(year=year)))
results = doc.select("div.content table tr")
for index, result in enumerate(results):
if not index:
# Skip the header row
continue
report = report_from(result, report_type="audit", year_range=year_range)
if report:
inspector.save_report(report)
# Pull the FOIA reports
doc = BeautifulSoup(utils.download(FOIA_REPORTS_URL))
results = doc.select("div.content table tr")
for index, result in enumerate(results):
if not index:
# Skip the header row
continue
report = report_from(result, report_type="other", year_range=year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.select("div.content a")
for result in results:
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options)
# Pull the audit reports
for year in year_range:
url = audit_report_url(year)
if url:
parse_result_from_js_url(url, "auditreports", year, year_range)
url = inspection_report_url(year)
if url:
parse_result_from_js_url(url, "iereports", year, year_range)
# Pull the congressional testimony
doc = BeautifulSoup(utils.download(CONGRESSIONAL_TESTIMONY_REPORTS_URL))
results = doc.findAll("ul", type='disc')[0].select("li")
for result in results:
report = congressional_testimony_report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.findAll("ul", type='disc')[0].select("li")
for result in results:
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
def urls_for(self):
only = self.options.get('topics')
if only: # if only...
only = set(only.split(','))
only = [(o, TOPIC_TO_REPORT_TYPE[o]) if o in TOPIC_TO_REPORT_TYPE else o
for o in only]
yield from self.urls_for_topics(only)
# If there are topics selected, ONLY yield URLs for those.
return
# First yield the URLs for the topics that are tangential to the main
# Calendar Year reports.
yield from self.urls_for_topics(ADDITIONAL_TOPICS)
# Not getting reports from specific topics, iterate over all Calendar Year
# reports.
page = BeautifulSoup(utils.download(BASE_URL))
# Iterate over each "Calendar Year XXXX" link
for li in page.select('.field-items li'):
md = RE_CALENDAR_YEAR.search(li.text)
if md:
cur_year = int(md.group(1))
if cur_year >= self.year_range[0] and cur_year <= self.year_range[-1]:
href = li.select('a')[0]['href']
next_url = urljoin(BASE_URL, href)
# The first page of reports is yielded.
yield next_url
# Next, read all the pagination links for the page and yield those. So
# far, I haven't seen a page that doesn't have all of the following
# pages enumerated.
next_page = BeautifulSoup(utils.download(next_url))
for link in next_page.select('li.pager-item a'):
yield urljoin(BASE_URL, link['href'])
def run(options):
year_range = inspector.year_range(options, archive)
doc = BeautifulSoup(utils.download(REPORTS_URL))
# Pull the semiannual reports
semiannul_results = doc.select("#AnnualManagementReports select")[0]
for result in semiannul_results.select("option"):
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the special reports
special_report_table = doc.find("table", attrs={"bordercolor": "#808080"})
for index, result in enumerate(special_report_table.select("tr")):
if not index:
# Skip the header row
continue
report = report_from(result, REPORTS_URL, report_type='other', year_range=year_range)
if report:
inspector.save_report(report)
# Pull the audit reports
for year in year_range:
if year < 2001: # The oldest fiscal year page available
continue
year_url = AUDIT_REPORTS_URL.format(year=year)
doc = BeautifulSoup(utils.download(year_url))
for index, result in enumerate(doc.select("#main table tr")):
if not index:
# Skip the header row
continue
report = report_from(result, year_url, report_type='audit', year_range=year_range)
if report:
inspector.save_report(report)
def extract_reports_for_oei(year_range):
topic_name = TOPIC_NAMES["OE"]
topic_url = TOPIC_TO_URL["OE"]
root_body = utils.download(topic_url)
root_doc = BeautifulSoup(root_body)
letter_urls = set()
for link in root_doc.select("#leftContentInterior li a"):
absolute_url = urljoin(topic_url, link['href'])
absolute_url = strip_url_fragment(absolute_url)
letter_urls.add(absolute_url)
if not letter_urls:
raise inspector.NoReportsFoundError("HHS (OEI first pass)")
all_results_links = {}
all_results_unreleased = []
for letter_url in letter_urls:
letter_body = utils.download(letter_url)
letter_doc = BeautifulSoup(letter_body)
results = letter_doc.select("#leftContentInterior ul li")
if not results:
raise inspector.NoReportsFoundError("HHS (OEI %s)" % letter_url)
for result in results:
if 'crossref' in result.parent.parent.attrs.get('class', []):
continue
if result.parent.parent.attrs.get('id') == 'related':
continue
node = result
while node and node.name != "h2":
node = node.previous
if node and node.name == "h2":
subtopic_name = str(node.text)
else:
subtopic_name = "(unknown)"
links = result.findAll("a")
if len(links) == 0:
result.extract()
all_results_unreleased.append([result, subtopic_name])
else:
url = links[0].get("href")
if url not in all_results_links:
result.extract()
all_results_links[url] = [result, subtopic_name]
else:
existing_result = all_results_links[url][0]
for temp in result.contents:
temp.extract()
existing_result.append(temp)
all_results_links[url][1] = "%s, %s" % (all_results_links[url][1], subtopic_name)
subtopic_url = TOPIC_TO_URL["OE"]
for result, subtopic_name in itertools.chain(all_results_links.values(), all_results_unreleased):
report = report_from(result, year_range, topic_name, subtopic_url, subtopic_name)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
doc = BeautifulSoup(utils.download(AUDIT_REPORTS_URL))
results = doc.select("td.text table tr")
if not results:
raise inspector.NoReportsFoundError("National Science Foundation (audit reports")
for result in results:
# ignore divider lines
if result.select("img"): continue
report = report_from(result, report_type='audit', year_range=year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
doc = BeautifulSoup(utils.download(SEMIANNUAL_REPORTS_URL))
results = doc.select("td.text table tr")
if not results:
raise inspector.NoReportsFoundError("National Science Foundation (semiannual reports)")
for result in results:
if not result.text.strip():
continue
report = semiannual_report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the case reports
response = utils.scraper.post(
url=CASE_REPORTS_URL,
data=CASE_REPORTS_DATA,
)
doc = BeautifulSoup(response.content)
results = doc.select("td.text table tr")
if not results:
raise inspector.NoReportsFoundError("National Science Foundation (case reports)")
for index, result in enumerate(results):
if not index or not result.text.strip(): # Skip the header row and empty rows
continue
report = case_report_from(result, CASE_REPORTS_URL, year_range)
if report:
inspector.save_report(report)
# Pull the testimony
doc = BeautifulSoup(utils.download(TESTIMONY_REPORTS_URL))
results = doc.select("td.text table tr")
if not results:
raise inspector.NoReportsFoundError("National Science Foundation (testimony)")
for result in results:
if not result.text.strip():
continue
report = report_from(result, report_type='testimony', year_range=year_range)
if report:
inspector.save_report(report)
def handle_scanner_args(args, opts) -> Tuple[dict, list]:
"""
--analytics: file path or URL to a CSV of participating domains.
This function also handles checking for the existence of the file,
downloading it succesfully, and reading the file in order to populate the
list of analytics domains.
"""
parser = scan_utils.ArgumentParser(prefix_chars="--")
parser.add_argument("--analytics", nargs=1, required=True)
parsed, unknown = parser.parse_known_args(args)
dicted = vars(parsed)
should_be_single = ["analytics"]
dicted = scan_utils.make_values_single(dicted, should_be_single)
resource = dicted.get("analytics")
if not resource.endswith(".csv"):
no_csv = "".join([
"--analytics should be the file path or URL to a CSV of participating",
" domains and end with .csv, which '%s' does not" % resource
])
logging.error(no_csv)
raise argparse.ArgumentTypeError(no_csv)
try:
parsed_url = urlparse(resource)
except:
raise
if parsed_url.scheme and parsed_url.scheme in ("http", "https"):
analytics_path = Path(opts["_"]["cache_dir"], "analytics.csv").resolve()
try:
utils.download(resource, str(analytics_path))
except:
logging.error(utils.format_last_exception())
no_csv = "--analytics URL %s not downloaded successfully." % resource
logging.error(no_csv)
raise argparse.ArgumentTypeError(no_csv)
else:
if not os.path.exists(resource):
no_csv = "--analytics file %s not found." % resource
logging.error(no_csv)
raise FileNotFoundError(no_csv)
else:
analytics_path = resource
analytics_domains = utils.load_domains(analytics_path)
dicted["analytics_domains"] = analytics_domains
del dicted["analytics"]
return (dicted, unknown)
def run(options):
year_range = inspector.year_range(options)
only_id = options.get('report_id')
print("## Downloading reports from %i to %i" % (year_range[0], year_range[-1]))
url = url_for()
body = utils.download(url)
doc = BeautifulSoup(body)
results = doc.select("section")
for result in results:
try:
year = int(result.get("title"))
# check that the fetched year is in the range
if year not in year_range:
continue
print("## Downloading year %i " % year)
except ValueError:
continue
# gets each table entry and sends generates a report from it
listings = result.div.table.tbody.contents
for item in listings:
if type(item) is not bs4.element.Tag:
continue
report = report_from(item)
# can limit it to just one report, for debugging convenience
if only_id and only_id != report['report_id']:
continue
inspector.save_report(report)
def fetch_from_landing_page(self, landing_url):
"""Returns a tuple of (pdf_link, summary_text, is_unreleased)."""
unreleased = False
page = BeautifulSoup(utils.download(landing_url))
summary = None
field_items = page.select('.field-items')
if field_items:
text = [node.strip() for node in field_items[0].findAll(text=True)]
summary = '\n\n'.join(text).strip()
if not summary:
logging.info('\tno summary text found')
if (summary and (RE_NOT_AVAILABLE.search(summary)
or RE_NOT_AVAILABLE_2.search(summary)
or RE_NOT_AVAILABLE_3.search(summary)
or RE_NOT_AVAILABLE_4.search(summary)
or RE_CLASSIFIED.search(summary))):
unreleased = True
report_url = None
pdf_link = page.select('.file a')
if not pdf_link:
logging.warn('No pdf link found on page: {0}'.format(landing_url))
else:
report_url = pdf_link[0]['href']
return report_url, summary, unreleased
def run(options):
year_range = inspector.year_range(options, archive)
report_flag = False
# Pull the table of reports for each year
for year in year_range:
url = url_for_year(year)
html = utils.download(url, scraper_slug="osc")
if html is None:
if year == max(year_range):
continue
else:
raise Exception("Couldn't fetch reports page {}".format(url))
# spaces appear as   and \u200b .... fix that now
html = html.replace(' ', ' ').replace('\u200b', ' ').replace('\u00a0', ' ').replace('\r', '').replace('\n', '')
doc = BeautifulSoup(html, "lxml")
OUTCOME_CODES = generate_outcome_codes(doc)
keys_used = [] # a few reports appear multiple times... ignore them the second time if they appear more than once
results = doc.findAll("table")[1].tbody.findAll('tr') # no ids on the tables, but it's the second one
for result in results:
reports = report_from(result, year, year_range, url, OUTCOME_CODES)
for report in reports:
if report['report_id'] not in keys_used:
inspector.save_report(report)
keys_used.append(report['report_id'])
report_flag = True
if not report_flag:
raise inspector.NoReportsFoundError("OSC")
def run(options):
year_range = inspector.year_range(options)
topics = options.get('topics')
if topics:
topics = topics.split(",")
else:
topics = TOPIC_TO_URL.keys()
for topic in topics:
topic_url = TOPIC_TO_URL[topic]
body = utils.download(topic_url)
doc = BeautifulSoup(body)
try:
year_results = doc.select("#Listing")[0]
results = [x for x in year_results.select("ul li ul li")]
except IndexError:
try:
all_results = doc.select("#bodyholder")[0]
results = [x for x in all_results.select("ul li")]
except IndexError:
results = doc.select("table ul li")
# Sometimes multiple reports are listed under the same datetime element.
# We store which published datetime we saw last so that the next report
# can use if if we are unable to find another published time.
last_published_on = None
for result in results:
report, last_published_on = report_from(result, topic_url, topic, year_range, last_published_on)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options)
for page_url in URLS:
done = False
body = utils.download(page_url)
doc = BeautifulSoup(body)
maincontent = doc.select("div#CS_Element_eximpagemaincontent")[0]
all_p = maincontent.find_all("p")
for p in all_p:
for all_text, link_text, link_url in recurse_tree(p, False):
if link_url == None:
continue
if link_url.startswith("mailto:"):
continue
if page_url == WHATS_NEW_URL and link_url == "/oig/whats-new-archive.cfm":
# end of page
done = True
break
if link_url.startswith("https://public.govdelivery.com/"):
continue
for index_url in URLS:
if index_url.find(link_url) != -1:
continue
year = DATE_RE.search(all_text).group(3)
if int(year) not in year_range:
continue
report = report_from(all_text, link_text, link_url, page_url)
inspector.save_report(report)
if done: break
def run(options):
year_range = inspector.year_range(options, archive)
component = options.get('component')
if component:
components = [component]
else:
components = list(COMPONENTS.keys())
report_id = options.get('report_id')
limit = int(options.get('limit', 0))
all_audit_reports = {}
for component in components:
logging.info("## Fetching reports for component %s" % component)
url = url_for(options, component)
body = utils.download(url)
doc = BeautifulSoup(body)
results = doc.select("table.contentpaneopen table[border=1] tr")
# accept only trs that look like body tr's (no 'align' attribute)
# note: HTML is very inconsistent. cannot rely on thead or tbody
results = [x for x in results if x.get('align') is None]
if not results:
raise inspector.NoReportsFoundError("DHS (%s)" % component)
count = 0
for result in results:
report = report_from(result, component, url)
if not report:
continue
if report_id and (report_id != report['report_id']):
continue
if inspector.year_from(report) not in year_range:
# logging.info("[%s] Skipping, not in requested range." % report['report_id'])
continue
key = (report["report_id"], report["title"])
if key in all_audit_reports:
all_audit_reports[key]["agency"] = all_audit_reports[key]["agency"] + \
", " + report["agency"]
all_audit_reports[key]["agency_name"] = \
all_audit_reports[key]["agency_name"] + ", " + \
report["agency_name"]
else:
all_audit_reports[key] = report
count += 1
if limit and (count >= limit):
break
logging.info("## Fetched %i reports for component %s\n\n" % (count, component))
for report in all_audit_reports.values():
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options)
doc = BeautifulSoup(utils.download(REPORTS_URL))
# Pull the audit reports
audit_header = doc.find("a", attrs={"name": 'Audit Reports'})
audit_list1 = audit_header.find_next("ul").select("li")
# They have two separate uls for these reports. See note to the IG web team.
audit_list2 = audit_header.find_next("ul").find_next("ul").select("li")
results = audit_list1 + audit_list2
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the inspection reports
inspections_header = doc.find("a", attrs={"name": 'Inspection Reports'})
results = inspections_header.find_next("ul").select("li")
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
semiannual_header = doc.find("a", attrs={"name": 'Semiannual Reports'})
results = semiannual_header.find_next("ul").select("li")
for result in results:
report = report_from(result, year_range, title_prefix="Semiannual Report - ")
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Find the number of pages to iterate
doc = BeautifulSoup(utils.download(REPORTS_URL))
page_count_text = doc.select("div.AspNet-GridView-Pagination")[0].text
page_count = int(re.search("Page 1 of (\d+)", page_count_text).groups()[0])
# Iterate over those pages
for page in range(1, page_count + 1):
response = utils.scraper.post(
REPORTS_URL,
data={
"__EVENTTARGET": "ctl00$ctl00$MainContent$NavTreeSubContent$sv$GridViewSummary",
"__EVENTARGUMENT": "Page${page_number}".format(page_number=page),
},
cookies=COOKIES,
)
doc = BeautifulSoup(response.content)
results = doc.select("div.AspNet-GridView table tr")
if not results:
break
for index, result in enumerate(results):
if not index:
# Skip the header row
continue
report = report_from(result, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options)
pages = options.get('pages', ALL_PAGES)
max_page = None
for page in range(1, (int(pages) + 1)):
if max_page and (page > max_page):
print("End of pages!")
break
print("## Downloading page %i" % page)
url = url_for(options, page)
body = utils.download(url)
doc = BeautifulSoup(body)
max_page = last_page_for(doc)
results = doc.select(".views-row")
for result in results:
report = report_from(result)
# inefficient enforcement of --year arg, USPS doesn't support it server-side
# TODO: change to published_on.year once it's a datetime
if inspector.year_from(report) not in year_range:
print("[%s] Skipping report, not in requested range." % report['report_id'])
continue
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options)
max_pages = int(options.get('pages', 1))
for year in year_range:
page = 1
done = False
while not done:
url = url_for(options, page, year)
body = utils.download(url)
doc = BeautifulSoup(body)
next_page = page + 1
found_next_page = False
page_links = doc.select("li.pager-item a.active")
for page_link in page_links:
if page_link.text == str(next_page):
found_next_page = True
break
if not found_next_page:
done = True
if next_page > max_pages:
done = True
results = doc.select("table.views-table > tbody > tr")
for result in results:
report = report_from(result)
inspector.save_report(report)
page = next_page
if not done:
print('Moving to next page (%d)' % page)
def run(options):
year_range = inspector.year_range(options, archive)
topics = options.get('topics')
if topics:
topics = topics.split(",")
else:
topics = TOPIC_TO_URL.keys()
all_reports = {}
for topic in topics:
year_urls = urls_for(year_range, topic)
for year_url in year_urls:
logging.debug("Scraping %s" % year_url)
body = utils.download(year_url)
doc = BeautifulSoup(body)
if not doc.select(".view-business-areas"):
raise inspector.NoReportsFoundError("DOT (%s)" % topic)
results = doc.select(".view-business-areas .views-row")
for result in results:
report = report_from(result, year_range, topic, options)
if report:
report_id = report["report_id"]
if report_id in all_reports:
all_reports[report_id]["topic"] = all_reports[report_id]["topic"] \
+ ", " + topic
else:
all_reports[report_id] = report
for report in all_reports.values():
inspector.save_report(report)
def fetch_from_landing_page(landing_url):
"""Returns a tuple of (pdf_link, summary_text)."""
add_pdf = False
body = utils.download(landing_url)
page = BeautifulSoup(body)
link = page.find('a', text=RE_PDF_LINK_TEXT, href=RE_PDF_HREF)
if not link:
link = page.find('a', text=RE_PDF_CLICK_TEXT, href=RE_PDF_HREF)
if not link:
link = page.find('a', text=RE_PDF_SARC_TEXT, href=RE_PDF_HREF)
# cases where .pdf is left off, ugh, e.g.
# http://www.dodig.mil/pubs/report_summary.cfm?id=849
if not link:
link = page.find('a', text=RE_PDF_LINK_TEXT, href=RE_BACKUP_PDF_HREF)
add_pdf = True
href = link['href'].strip() if link else None
if href and add_pdf:
href = href + ".pdf"
summary = None
text_tr = page.select('tr[valign="top"] td')
if text_tr:
text = [node.strip() for node in text_tr[0].findAll(text=True)]
summary = '\n\n'.join(text)
if not summary:
logging.info('\tno summary text found')
return (href, summary)
def extract_reports_for_subtopic(subtopic_url, year_range, topic, subtopic=None):
if subtopic_url.startswith("http://httphttp://"):
# See notes to IG's web team
subtopic_url = subtopic_url.replace("http://http", "")
body = utils.download(subtopic_url)
doc = BeautifulSoup(body)
results = doc.select("#body-row02-col02andcol03 a")
if not results:
results = doc.select("#body-row02-col01andcol02andcol03 a")
if not results and "There are currently no reports in this category" not in doc.text:
raise AssertionError("No report links found for %s" % subtopic_url)
topic_name = TOPIC_NAMES[topic]
# Broadcasting Board of Governors is a fully independent agency
if topic == 'BBG' or subtopic == 'Broadcasting Board of Governors':
agency = 'bbg'
else:
agency = 'state'
for result in results:
report = report_from(result, year_range, agency, topic_name, subtopic)
if report:
inspector.save_report(report)
def semiannual_report_from(result, year_range):
link = result.find("a")
title = link.text
# Parse the report title. Ex:
# 'OIG Semiannual Report to the Congress: October 1, 2013 - March 31, 2014 (incl. MCC)'
published_on_text = title.split("-")[-1].split("–")[-1].split("(")[0].strip()
published_on_text = published_on_text.replace("September 31", "September 30") # See note to IG Web team
published_on = datetime.datetime.strptime(published_on_text, '%B %d, %Y')
if published_on.year not in year_range:
logging.debug("[%s] Skipping, not in requested range." % title)
return
landing_url = urljoin(SEMIANNUAL_REPORTS_URL, link.get('href'))
landing_page = BeautifulSoup(utils.download(landing_url))
report_url = landing_page.select("div.filefield-file a")[0].get('href')
report_filename = report_url.split("/")[-1]
report_id, _ = os.path.splitext(report_filename)
report = {
'inspector': "usaid",
'inspector_url': "https://oig.usaid.gov",
'agency': "usaid",
'agency_name': "Agency For International Development",
'type': 'semiannual_report',
'report_id': report_id,
'url': report_url,
'title': title,
'published_on': datetime.datetime.strftime(published_on, "%Y-%m-%d"),
}
return report
def run(options):
year_range = inspector.year_range(options, archive)
pages = options.get('pages', ALL_PAGES)
# default to starting at page 1
begin = int(options.get('begin', 1))
max_page = None
for page in range(begin, (int(pages) + 1)):
if max_page and (page > max_page):
logging.debug("End of pages!")
break
logging.debug("## Downloading page %i" % page)
url = url_for(options, page)
body = utils.download(url)
doc = BeautifulSoup(body)
# When the USPS restores their page controls, we can use this again,
# which saves one network call each time.
max_page = last_page_for(doc)
results = doc.select(".views-row")
for result in results:
report = report_from(result)
# inefficient enforcement of --year arg, USPS doesn't support it server-side
# TODO: change to published_on.year once it's a datetime
if inspector.year_from(report) not in year_range:
logging.warn("[%s] Skipping report, not in requested range." % report['report_id'])
continue
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the reports
for report_type, report_url in REPORT_URLS.items():
doc = BeautifulSoup(utils.download(report_url))
results = doc.select("td.mainInner div.ms-WPBody li")
if not results:
raise inspector.NoReportsFoundError("SIGTARP (%s)" % report_url)
for result in results:
report = report_from(result, report_type, year_range)
if report:
inspector.save_report(report)
def beautifulsoup_from_url(url):
body = utils.download(url)
if body is None: return None
doc = BeautifulSoup(body)
# Some of the pages will return meta refreshes
if doc.find("meta") and doc.find("meta").attrs.get(
'http-equiv') == 'REFRESH':
redirect_url = urljoin(
url,
doc.find("meta").attrs['content'].split("url=")[1])
return beautifulsoup_from_url(redirect_url)
else:
return doc
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the reports
doc = BeautifulSoup(utils.download(REPORTS_URL))
results = doc.find("table", {"cellpadding": "5"}).select("tr")
if not results:
raise inspector.NoReportsFoundError("FDIC")
for index, result in enumerate(results):
if index < 3 or not result.text.strip():
# The first three rows are headers
continue
report = report_from(result, year_range)
if report:
inspector.save_report(report)
def parse_result_from_js_url(url, format_slug, year, year_range, report_type):
"""
Given a link to a javascript file that has report data, add all of the reports
"""
body = utils.download(url, scraper_slug="tigta")
# Pulling out javascript array values that look like:
# arrid[0]=new AR("200720002","Stronger Management Oversight Is Required to Ensure Valuable Systems Modernization Expertise Is Received From the Federally Funded Research and Development Center Contractor","20061020","01",2,0,0,0);
# Look in https://www.treasury.gov/tigta/oa_auditreports_fy14.js for some more examples.
results = re.findall('arrid\[\d+\]=new AR\((.*)\);', body)
for result in results:
report = report_from(result, format_slug, year, year_range,
report_type)
if report:
inspector.save_report(report)
def get_pagination_urls(page):
"""Find the pagination links on the page and yield them all.
This method recursively downloads new pages in the case that there are more
than 10.
"""
for link in page.select('a'):
if 'href' not in link:
continue
if link['href'].startswith('?') and RE_DIGITS.match(link.text):
yield BASE_URL + link['href']
elif link['href'].startswith('/pubs') and RE_NEXT_10.search(link.text):
new_url = urljoin(BASE_URL, link['href'])
page = BeautifulSoup(utils.download(new_url))
for link in get_pagination_urls(page):
yield link
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the reports
for report_type, url in REPORT_TYPE_MAP.items():
doc = BeautifulSoup(utils.download(url))
results = doc.select("div#content div#contentMain ul li.pdf")
if not results:
raise inspector.NoReportsFoundError("CPB (%s)" % url)
for result in results:
if not result.find('a'):
# Skip unlinked PDF's
continue
report = report_from(result, url, report_type, year_range)
if report:
inspector.save_report(report)
def get_subtopic_map(topic_url):
body = utils.download(topic_url)
doc = BeautifulSoup(body)
subtopic_map = {}
for link in doc.select("#leftContentInterior li a"):
absolute_url = urljoin(topic_url, link['href'])
absolute_url = strip_url_fragment(absolute_url)
# Only add new URLs
if absolute_url not in subtopic_map.values():
subtopic_map[link.text] = absolute_url
if not subtopic_map:
raise inspector.NoReportsFoundError("OEI (subtopics)")
return subtopic_map
def reports_from_page(url_format, page, report_type, year_range, year=''):
url = url_format.format(page=page, year=year)
doc = BeautifulSoup(utils.download(url))
results = doc.select("td.views-field")
if not results:
results = doc.select("div.views-row")
if not results:
return False
for result in results:
if not result.text.strip():
# Skip empty rows
continue
report = report_from(result, report_type, year_range)
if report:
inspector.save_report(report)
return True
def run(options):
year_range = inspector.year_range(options, archive)
doc = BeautifulSoup(utils.download(REPORTS_URL))
results = None
for section in doc.find_all("section"):
if section.h4 and section.h4.text.strip() == "Publications":
results = section.find_all("a")
break
if not results:
raise inspector.NoReportsFoundError("Denali Commission")
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
def semiannual_report_from(result, year_range):
# This will look like "toggleReport('SARC-47-49');" and we want to pull out
# the SARC-47-49
report_id_javascript = result.get('onclick')
report_id = re.search("'(.*)'", report_id_javascript).groups()[0]
landing_url = "http://oig.pbgc.gov/sarc/{report_id}.html".format(
report_id=report_id)
landing_page = BeautifulSoup(utils.download(landing_url))
title = " ".join(landing_page.select("h3")[0].text.split())
relative_report_url = landing_page.find(
"a", text="Read Full Report").get('href')
# The relative report urls try to go up a level too many. Most browsers seem
# to just ignore this so we will too.
relative_report_url = relative_report_url.replace("../", "", 1)
report_url = urljoin(SEMIANNUAL_REPORTS_URL, relative_report_url)
# There is probably a way to be a bit smarter about this
summary = landing_page.text.strip()
published_on_text = title.rsplit("-")[-1].rsplit("through")[-1].replace(
".", "").strip()
published_on = datetime.datetime.strptime(published_on_text, '%B %d, %Y')
if published_on.year not in year_range:
logging.debug("[%s] Skipping, not in requested range." % title)
return
report = {
'inspector': "pbgc",
'inspector_url': "http://oig.pbgc.gov",
'agency': "pbgc",
'agency_name': "Pension Benefit Guaranty Corporation",
'type': 'semiannual_report',
'report_id': report_id,
'url': report_url,
'title': title,
'published_on': datetime.datetime.strftime(published_on, "%Y-%m-%d"),
}
if summary:
report['summary'] = summary
if landing_url:
report['landing_url'] = landing_url
return report
def extract_from_release_page(landing_url):
doc = BeautifulSoup(utils.download(landing_url))
main = doc.select("#main #lefSide")[0]
url_elem = main.select("div")[2].select("a")
if url_elem:
url = urljoin(landing_url, url_elem[0]['href'])
else:
url = None
summary = ""
for p in main.select("p"):
summary += p.text + "\n\n"
# will only be used if the title isn't present on the listing
title = main.select("h2")[0].text.strip()
return (url, summary.strip(), title)
def run(options):
year_range = inspector.year_range(options, archive)
for index in INDEX_URLS:
report_count = 0
for year in year_range:
url = url_for(options, index, year)
body = utils.download(url)
doc = BeautifulSoup(body)
results = doc.select("div.view-content div.views-row")
for result in results:
report = report_from(result)
inspector.save_report(report)
report_count = report_count + 1
if report_count == 0:
raise inspector.NoReportsFoundError("Amtrak (%s)" % index.split("/")[-1])
def run(options):
year_range = inspector.year_range(options, archive)
last_page = options.get("end")
start = int(options.get("start", 1))
# Pull the reports
for (reports_page, report_type) in REPORTS_URLS:
page = start
last_page = options.get("end") # reset for each area
while True:
url = url_for(reports_page, page)
doc = BeautifulSoup(utils.download(url))
if last_page is None:
last_page = last_page_from(doc)
if report_type == "case":
results = doc.select("div#main div.grayBox2")
else:
results = doc.select("div#main div.whiteBox")
if results:
for result in results:
report = report_from(result, reports_page, report_type,
year_range)
if report:
inspector.save_report(report)
elif report_type != "case":
raise inspector.NoReportsFoundError("CNCS (%s)" % url)
# closed cases have broken pagination (p6, 7, 8 missing) so ignore
else:
pass
if int(page) >= int(last_page):
break
else:
page += 1
# one hardcoded peer review, just always do it
inspector.save_report(do_peer_review())
def run(options):
year_range = inspector.year_range(options, archive)
only_id = options.get('report_id')
logging.info("## Downloading reports from %i to %i" %
(year_range[0], year_range[-1]))
url = url_for()
body = utils.download(url)
doc = BeautifulSoup(body)
results = doc.select("section")
if not results:
raise inspector.NoReportsFoundError("OPM")
for result in results:
try:
year = int(result.get("title"))
# check that the fetched year is in the range
if year not in year_range:
continue
logging.info("## Downloading year %i " % year)
except ValueError:
continue
# gets each table entry and sends generates a report from it
listings = result.div.table.tbody.contents
for item in listings:
if type(item) is not bs4.element.Tag:
continue
report = report_from(item)
if report['report_id'] in BLACKLIST:
logging.warn(
"Skipping downed report: remember to report this and get it fixed!"
)
continue
# can limit it to just one report, for debugging convenience
if only_id and only_id != report['report_id']:
continue
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
only = options.get('topics')
if only:
only = set(only.split(','))
else:
only = ALL_TOPIC_AREAS
index_body = utils.download(BASE_URL)
current_year = None
index = BeautifulSoup(index_body)
tables = index.select('table.style1')
if not tables:
raise inspector.NoReportsFoundException("EPA")
for table in tables:
trs = table.select('tr')
for tr in trs:
tds = tr.select('td')
if len(tds) < 8:
if len(tds) == 1:
# Large column that indicates year
col_links = tds[0].select('a')
if len(col_links) == 1:
col_text = col_links[0].text
if RE_YEAR.match(col_text):
current_year = col_text
continue
published_on_dt = datetime.datetime.strptime(
tds[6].text, '%m/%d/%Y')
if published_on_dt.year not in year_range:
continue
topic_areas = set(tds[7].text.split(', '))
if not len(topic_areas.intersection(only)):
continue
report = report_from(tds, published_on_dt, current_year)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
topics = options.get('topics')
if topics:
topics = topics.split(",")
else:
topics = TOPIC_TO_URL.keys()
for topic in topics:
topic_url = TOPIC_TO_URL[topic]
body = utils.download(topic_url)
doc = BeautifulSoup(body)
try:
year_results = doc.select("#Listing")[0]
results = [x for x in year_results.select("ul li ul li")]
except IndexError:
try:
all_results = doc.select("#bodyholder")[0]
results = [x for x in all_results.select("ul li")]
except IndexError:
results = doc.select("table ul li")
if not results:
raise inspector.NoReportsFoundError("SEC (%s)" % topic)
# Sometimes multiple reports are listed under the same datetime element.
# We store which published datetime we saw last so that the next report
# can use if if we are unable to find another published time.
last_published_on = None
for result in results:
report, last_published_on = report_from(result, topic_url, topic,
year_range,
last_published_on)
if report:
inspector.save_report(report)
for canned_report in CANNED_REPORTS:
report_datetime = datetime.datetime.strptime(
canned_report["published_on"], "%Y-%m-%d")
if report_datetime.year in year_range:
add_common_fields(canned_report)
inspector.save_report(canned_report)
def urls_for(options, only):
year_range = inspector.year_range(options, archive)
for office in only:
# there's always a first year, and it defaults to current year
params = {}
params['searchdate1'] = '01/01/%s' % year_range[0]
params['searchdate2'] = '12/31/%s' % year_range[-1] # could be the same year
params['office'] = OFFICES[office]
params['sort'] = 'report_number'
params['order'] = 'desc'
query_string = urlencode(params)
url = '{0}?{1}'.format(BASE_URL, query_string)
yield url
body = utils.download(url)
page = BeautifulSoup(body)
for url in get_pagination_urls(page):
yield url
def run(options):
only = options.get('topics')
if only:
only = set(only.split(','))
else:
# Default to all offices, whee!
only = list(OFFICES.keys())
for url in urls_for(options, only):
body = utils.download(url)
page = BeautifulSoup(body)
report_table = page.select('table[summary~="reports"]')[0]
for tr in report_table.select('tr')[1:]:
tds = tr.select('td')
if len(tds) == 1:
# Page has no reports, simply a "No Data" indication for these dates.
break
report = report_from(tds, options)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
doc = BeautifulSoup(utils.download(REPORTS_URL))
# Pull the audit reports
audit_header = doc.find("a", attrs={"name": 'Audit Reports'})
audit_list1 = audit_header.find_next("ul").select("li")
# They have two separate uls for these reports. See note to the IG web team.
audit_list2 = audit_header.find_next("ul").find_next("ul").select("li")
results = audit_list1 + audit_list2
if not results:
raise inspector.NoReportsFoundError("FEC (audit reports)")
for result in results:
report = report_from(result, year_range, report_type='audit')
if report:
inspector.save_report(report)
# Pull the inspection reports
inspections_header = doc.find("a", attrs={"name": 'Inspection Reports'})
results = inspections_header.find_next("ul").select("li")
if not results:
raise inspector.NoReportsFoundError("FEC (inspection reports)")
for result in results:
report = report_from(result, year_range, report_type='inspection')
if report:
inspector.save_report(report)
# Pull the semiannual reports
semiannual_header = doc.find("a", attrs={"name": 'Semiannual Reports'})
results = semiannual_header.find_next("ul").select("li")
if not results:
raise inspector.NoReportsFoundError("FEC (semiannual reports)")
for result in results:
report = report_from(result,
year_range,
report_type='semiannual_report',
title_prefix="Semiannual Report - ")
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the reports
for report_type, url in REPORT_URLS.items():
doc = BeautifulSoup(utils.download(url))
results = doc.select("div.section1 div.ltext > table tr")
if not results:
results = doc.select(
"td.three-col-layout-middle div.ltext > table tr")
if not results:
raise inspector.NoReportsFoundError(
"Government Publishing Office (%s)" % url)
for result in results:
if (not result.text.strip() or result.find("th")
or result.find("strong")
or result.contents[1].text in HEADER_TITLES):
# Skip header rows
continue
report = report_from(result, url, report_type, year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
max_pages = options.get('pages', None)
if max_pages:
max_pages = int(max_pages)
for year in year_range:
page = 1
done = False
while not done:
url = url_for(options, page, year)
body = utils.download(url)
doc = BeautifulSoup(body)
next_page = page + 1
found_next_page = False
page_links = doc.select("li.pager-item a.active")
for page_link in page_links:
if page_link.text == str(next_page):
found_next_page = True
break
if not found_next_page:
done = True
if max_pages and (next_page > max_pages):
done = True
results = doc.select("table.views-table > tbody > tr")
if not results:
raise inspector.NoReportsFoundError("Amtrak")
for result in results:
report = report_from(result)
inspector.save_report(report)
page = next_page
if not done:
logging.info('Moving to next page (%d)' % page)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
doc = BeautifulSoup(utils.download(AUDIT_REPORTS_URL))
headers = doc.select("p.Ptitle1")
if not headers:
raise inspector.NoReportsFoundError("ITC")
for header in headers:
year = int(header.text.strip())
results = header.findNextSibling("ul").select("li")
for result in results:
if not inspector.sanitize(result.text):
logging.debug("Skipping empty list item.")
continue
report = audit_report_from(year, result, AUDIT_REPORTS_URL,
year_range)
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
# Pull the audit reports
doc = BeautifulSoup(utils.download(REPORTS_URL))
results = doc.select("ul.text > ul > li")
if not results:
raise inspector.NoReportsFoundError("CFTC audit reports")
for result in results:
report = report_from(result, year_range)
if report:
inspector.save_report(report)
# Pull the semiannual reports
results = doc.select("ul.text td a")
if not results:
raise inspector.NoReportsFoundError("CFTC semiannual reports")
for result in results:
report = report_from(result,
year_range,
report_type="semiannual_report")
if report:
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
only_report_id = options.get('report_id')
# Pull the reports
for report_type, url in REPORT_URLS.items():
doc = BeautifulSoup(utils.download(url))
results = doc.select("div.field-item li")
if not results:
results = doc.select("div.field-item tr")
if not results:
raise inspector.NoReportsFoundError(
"National Endowment for the Arts (%s)" % report_type)
for result in results:
report = report_from(result, url, report_type, year_range)
if report:
# debugging convenience: can limit to single report
if only_report_id and (report['report_id'] != only_report_id):
continue
inspector.save_report(report)
def run(options):
year_range = inspector.year_range(options, archive)
report_flag = False
# Pull the table of reports for each year
for year in year_range:
url = url_for_year(year)
html = utils.download(url, scraper_slug="osc")
if html is None:
if year == max(year_range):
continue
else:
raise Exception("Couldn't fetch reports page {}".format(url))
# spaces appear as   and \u200b .... fix that now
html = html.replace(' ', ' ').replace('\u200b', ' ').replace(
'\u00a0', ' ').replace('\r', '').replace('\n', '')
doc = BeautifulSoup(html, "lxml")
OUTCOME_CODES = generate_outcome_codes(doc)
keys_used = [
] # a few reports appear multiple times... ignore them the second time if they appear more than once
results = doc.findAll("table")[1].tbody.findAll(
'tr') # no ids on the tables, but it's the second one
for result in results:
reports = report_from(result, year, year_range, url, OUTCOME_CODES)
for report in reports:
if report['report_id'] not in keys_used:
inspector.save_report(report)
keys_used.append(report['report_id'])
report_flag = True
if not report_flag:
raise inspector.NoReportsFoundError("OSC")
def run(options):
year_range = inspector.year_range(options, archive)
pages = options.get('pages', ALL_PAGES)
# default to starting at page 1
begin = int(options.get('begin', 1))
max_page = None
for page in range(begin, (int(pages) + 1)):
if max_page and (page > max_page):
logging.debug("End of pages!")
break
logging.debug("## Downloading page %i" % page)
url = url_for(options, page)
body = utils.download(url)
doc = BeautifulSoup(body)
# When the USPS restores their page controls, we can use this again,
# which saves one network call each time.
max_page = last_page_for(doc)
results = doc.select(".views-row")
if not results:
raise inspector.NoReportsFoundError("USPS")
for result in results:
report = report_from(result)
# inefficient enforcement of --year arg, USPS doesn't support it server-side
# TODO: change to published_on.year once it's a datetime
if inspector.year_from(report) not in year_range:
logging.warn("[%s] Skipping report, not in requested range." %
report['report_id'])
continue
inspector.save_report(report)
def report_from(result, report_type, year_range):
landing_page_link = result.find("a")
title = landing_page_link.text.strip()
landing_url = urljoin(BASE_REPORT_URL, landing_page_link.get('href'))
# Sometimes the last report on one page is also the first report on the next
# page. Here, we skip any duplicate landing pages we've already saved.
if landing_url in visited_landing_urls:
return
# This landing page is a duplicate of another one
if landing_url == "http://oig.ssa.gov/physical-security-office-disability-" \
"adjudication-and-reviews-headquarters-building-limited-0":
return
published_on_text = result.select(
"span.date-display-single")[0].text.strip()
published_on = datetime.datetime.strptime(published_on_text,
'%A, %B %d, %Y')
if published_on.year not in year_range:
logging.debug("[%s] Skipping, not in requested range." % title)
return
try:
report_id = result.select("span.field-data")[0].text.strip()
except IndexError:
report_id = landing_url.split("/")[-1]
# This report has the wrong report number entered
if landing_url == "http://oig.ssa.gov/audits-and-investigations/" \
"audit-reports/congressional-response-report-internet-claim-" \
"applications-0":
report_id = "A-07-10-20166"
landing_page = BeautifulSoup(utils.download(landing_url))
unreleased = False
if "Limited Distribution" in title:
unreleased = True
report_url = None
else:
try:
report_url = result.select("span.file a")[0].get('href')
except IndexError:
if not unreleased:
try:
report_url = landing_page.find(
"a", attrs={
"type": 'application/octet-stream;'
}).get('href')
except AttributeError:
unreleased = True
report_url = None
try:
summary = landing_page.select(
"div.field-type-text-with-summary")[0].text.strip()
except IndexError:
summary = None
file_type = None
if report_url:
_, extension = os.path.splitext(report_url)
if not extension:
file_type = 'html'
visited_landing_urls.add(landing_url)
report = {
'inspector': "ssa",
'inspector_url': "http://oig.ssa.gov",
'agency': "ssa",
'agency_name': "Social Security Administration",
'type': report_type,
'landing_url': landing_url,
'report_id': report_id,
'url': report_url,
'title': title,
'published_on': datetime.datetime.strftime(published_on, "%Y-%m-%d"),
}
if unreleased:
report['unreleased'] = unreleased
if file_type:
report['file_type'] = file_type
if summary:
report['summary'] = summary
return report
def beautifulsoup_from_url(url):
body = utils.download(url)
return BeautifulSoup(body)
def run(options):
year_range = inspector.year_range(options, archive)
published_on = None
for page_url in [
WHATS_NEW_URL, WHATS_NEW_ARCHIVE_URL,
SEMIANNUAL_REPORTS_AND_TESTIMONIES_URL
]:
body = utils.download(page_url)
doc = BeautifulSoup(body)
maincontent = doc.select("div#CS_Element_eximpagemaincontent")[0]
all_a = maincontent.find_all("a")
if not all_a:
raise inspector.NoReportsFoundException("Ex-Im Bank (%s)" %
page_url)
for a in all_a:
a_text = str(a.text)
if a_text.strip() == "":
continue
a_href = a.get("href")
if a_href.startswith("mailto:"):
continue
if a_href.startswith("https://public.govdelivery.com/"):
continue
if page_url == WHATS_NEW_URL and a_href == "/oig/whats-new-archive.cfm":
# end of page
break
if deduplicate_url(a_href):
continue
# Now, we want to grab all of the text associated with this link.
# If there is just one link inside of a paragraph tag, we can take the
# text contents of that paragraph tag. Otherwise, we use "previous" to
# grab all the text that comes before the link.
parent_p = a
while parent_p.name != "p":
parent_p = parent_p.parent
links_in_parent = parent_p.find_all("a")
links_in_parent = [link for link in links_in_parent \
if len(link.text.strip())]
links_in_parent = set(
[link.get("href") for link in links_in_parent])
if len(links_in_parent) == 1:
all_text = parent_p.text
else:
all_text = a_text
node = a.previous
while True:
if is_inside_link(node):
break
if isinstance(node, NavigableString):
all_text = node + all_text
node = node.previous
if not node:
break
if node == maincontent:
break
# Response letters don't get their own date heading -- keep date from
# last report and reuse in those cases
temp = DATE_RE.search(all_text)
if temp:
# For semiannual reports to congress, use the second date from the text
# Also, tack the date onto the report_id to disambiguate
if page_url == SEMIANNUAL_REPORTS_AND_TESTIMONIES_URL and a_text.strip(
).startswith('Semiannual Report to Congress'):
a_text = a_text.strip() + ' ' + temp.group(0) + ' - '
temp = DATE_RE.search(all_text, temp.end() + 1)
a_text = a_text + temp.group(0)
date_text = temp.group(0).replace('Sept ', 'Sep ')
try:
published_on = datetime.strptime(date_text, '%B %d, %Y')
except ValueError:
published_on = datetime.strptime(date_text, '%b %d, %Y')
if (published_on is None) or (published_on.year not in year_range):
continue
report = report_from(all_text, a_text, a_href, page_url,
published_on)
inspector.save_report(report)
for page_url in [PRESS_RELEASES_URL, PRESS_RELEASES_ARCHIVE_URL]:
done = False
body = utils.download(page_url)
doc = BeautifulSoup(body)
maincontent = doc.select("div#CS_Element_eximpagemaincontent")[0]
all_p = maincontent.find_all("p")
for p in all_p:
for all_text, link_text, link_url in recurse_tree(p, False):
if link_url == None:
continue
if link_url.startswith("mailto:"):
continue
if page_url == WHATS_NEW_URL and link_url == "/oig/whats-new-archive.cfm":
# end of page
done = True
break
if link_url.startswith("https://public.govdelivery.com/"):
continue
for index_url in URLS:
if index_url.find(link_url) != -1:
continue
if deduplicate_url(link_url):
continue
date_match = DATE_RE.search(all_text)
try:
date_text = date_match.group(0).replace('Sept ', 'Sep ')
published_on = datetime.strptime(date_text, '%B %d, %Y')
except ValueError:
published_on = datetime.strptime(date_text, '%b %d, %Y')
if published_on.year not in year_range:
continue
report = report_from(all_text, link_text, link_url, page_url,
published_on)
inspector.save_report(report)
if done: break
def fetch_from_landing_page(landing_url):
"""Returns a tuple of (pdf_link, summary_text)."""
add_pdf = False
skip = False
body = utils.download(landing_url)
page = BeautifulSoup(body)
report_tables = page.select('table[summary~="reports"]')
# in the rare case that doesn't work, have faith
if len(report_tables) == 0:
report_tables = page.select('table')
table = report_tables[0]
examine_text = table.text
maybe_unreleased = False
if RE_OFFICIAL.search(examine_text) or RE_CLASSIFIED.search(examine_text) or RE_FOIA.search(examine_text) or RE_AFGHANISTAN.search(examine_text) or RE_RESTRICTED.search(examine_text) or RE_INTEL.search(examine_text):
# 'Official use only' or 'Classified' materials don't have PDFs. Mark the
# report metadata appropriately.
maybe_unreleased = True
# two varieties of normal report link
link = page.find('a', text=RE_PDF_LINK_TEXT, href=RE_PDF_HREF)
if not link:
link = page.find('a', text=RE_PDF_CLICK_TEXT, href=RE_PDF_HREF)
# Semi annual reports to Congress
if not link:
link = page.find('a', text=RE_PDF_SARC_TEXT, href=RE_PDF_HREF)
# occurs for some multi-part reports, top/body/bottom
if not link:
link = page.find('a', text=RE_PDF_BODY_MAYBE, href=RE_PDF_HREF)
# cases where .pdf is left off, ugh, e.g.
# http://www.dodig.mil/pubs/report_summary.cfm?id=849
if not link:
link = page.find('a', text=RE_PDF_LINK_TEXT, href=RE_BACKUP_PDF_HREF)
if link:
add_pdf = True
# last resort, slow python-based check for tightest requirements
if not link:
link = page.find(pdf_test)
# before accepting *any* PDF, check for skippable offenses
if not link and (RE_EXTERNALLY_HOSTED.search(table.text) or RE_RESCINDED.search(table.text) or RE_RETRACTED.search(table.text) or RE_UNUSED.search(table.text)):
skip = True
# okay, I'll take *any* PDF
if not link:
link = table.find(any_pdf_test)
href = link['href'].strip() if link else None
if href and add_pdf:
href = href + ".pdf"
# some URLs have "/../" in the middle, and the redirects are trouble
if href:
href = href.replace("/../", "/")
summary = None
text_tr = page.select('tr[valign="top"] td')
if text_tr:
text = [node.strip() for node in text_tr[0].findAll(text=True)]
summary = '\n\n'.join(text)
if not summary:
logging.info('\tno summary text found')
return (href, summary, maybe_unreleased, skip)
