def annotate_page(scan, size, articles):
scan_px = scan.load()
log.debug("Image size: {} x {}".format(scan.size[0], scan.size[1]))
log.debug("Page size: {} x {}".format(int(size[0]), int(size[1])))
# the image size is not the same as the page size,
# so we must calculate how different it is.
fx = size[0] / scan.size[0]
fy = size[1] / scan.size[1]
log.debug("Scale factors: " + str((fx, fy)))
log.info("Annotating scan...")
# indicate elements
log.debug("Drawing the following elements:")
log.indent()
# color all articles
RED = (255, 0, 0)
for article in articles:
left = math.floor(article['start'][0] / fx)
right = math.ceil(article['end'][0] / fx)
top = math.floor(article['start'][1] / fy)
bottom = math.ceil(article['end'][1] / fy)
log.debug("Rectangle[{:>4} + {:>4}, {:>4} + {:>4}]".format(
left,
right - left,
bottom,
bottom - top,
))
# margins are marked in solid red
# top and bottom
for i in range(left, right):
scan_px[i, top] = RED
scan_px[i, bottom - 1] = RED
# left and right
for i in range(top, bottom):
scan_px[left, i] = RED
scan_px[right - 1, i] = RED
# color the inside of the block yellow
for i in range(left + 1, right):
for j in range(top + 1, bottom):
scan_px[i, j] = (min(255, scan_px[i, j][0] + 40),
min(255,
scan_px[i, j][1] + 40), scan_px[i, j][2])
log.dedent()
def handle_page(root):
# log.info("Parsing ALTO file...")
# in case server gives an HTTP 200 response with invalid data, try
# to detect it at parse time.
try:
alto = parsefile('ocr.xml')
except SAXParseException as e:
log.error("Invalid XML found. Maybe site is down?")
return
if alto is None:
log.warn("Non-conforming ALTO file found. Skipping.")
return
# show font info for page
log.debug("Fonts:")
log.indent()
for name, font in alto.fonts.items():
log.debug("{:<6}: {}".format(name, font))
log.dedent()
# log.info("Finding articles...")
find_articles(alto)
if '--annotate' in sys.argv and 'scan.jp2' in os.listdir():
log.info("Opening image file...")
scan = Image.open('scan.jp2').convert('RGB')
# annotate image with articles
with open('articles.json') as articles:
annotate_page(scan, alto.page.size, json.load(articles))
# save and finish
scan.save('annotated_scan.jpg')
scan.close()
# insert into database
try:
articleToDB.putArticlesInDatabase(commitIn=1, root=root)
except Exception as e:
# if database insertion fails, quit completely.
# have user restart process where it failed with
# `--start` option.
log.error("Database error: {}".format(e))
for file in os.listdir():
os.remove(file)
raise e
def wrap(*args, **kwargs):
log.message(start)
log.indent()
return_value = None
try:
return_value = func(*args, **kwargs)
except:
log.exception()
if check:
if not return_value:
log.error("Failure")
log.dedent()
return return_value
def fetch_batch(batch, start_issue):
metadata = {}
batch_url = CA_DOMAIN + "/batches/" + batch + ".json"
# fetch individual batch json
batch = poll_fetch(batch_url)
if batch is None:
log.warn("Could not fetch batch. Skipping.")
return
batch_json = batch.json()
log.info('Issues: ' + str(len(batch_json['issues'])))
log.info('Pages: ' + str(batch_json['page_count']))
log.indent()
for issue in batch_json['issues']:
metadata['newspaper'] = issue['title']['name']
metadata['publication_date'] = issue['date_issued']
if (batch_json['issues'].index(issue) + 1) < start_issue:
continue
# unit testing
#if (batch_json['issues'].index(issue)+1) > start_issue:
# return
log.info('Issue {}/{}: "{}", {}'.format(
batch_json['issues'].index(issue) + 1, len(batch_json['issues']),
metadata['newspaper'], metadata['publication_date']))
# fetch title json for publisher and location data
title = poll_fetch(issue['title']['url'])
if title is None:
log.warn("Could not fetch title information. Skipping.")
continue
title_json = title.json()
metadata['publisher'] = title_json['publisher']
place = title_json['place'][0].split('--')
metadata['state'] = place[0]
metadata['city'] = place[-1]
# fetch json for newspaper issue
issue = poll_fetch(issue['url'])
if issue is None:
log.warn("Could not fetch issue JSON. Skipping.")
continue
issue_json = issue.json()
# sort pages by sequence number, just in case
issue_json['pages'].sort(key=lambda page: page['sequence'])
log.info('Saving {} pages at ./{}'.format(
len(issue_json['pages']),
os.path.splitext(urlparse(issue_json['url']).path[1:])[0]))
log.indent()
for page in issue_json['pages']:
log.info("Processing page {} ...".format(page['sequence']))
# fetch json for single page
page = poll_fetch(page['url'])
if page is None:
log.warn("Could not fetch page JSON. Skipping.")
continue
page_json = page.json()
metadata['source_link'] = page_json['pdf']
log.debug("Metadata: {")
log.indent()
for key in metadata:
log.debug("'{}': '{}'".format(key, metadata[key]))
log.dedent()
log.debug("}")
# parse ocr url and get its path component only
path = os.path.dirname(urlparse(page_json['ocr']).path[1:])
root = os.getcwd()
os.makedirs(path, exist_ok=True)
os.chdir(path)
# save metadata.json; SHOULD work
with open('metadata.json', 'w') as file:
file.write(json.dumps(metadata, sort_keys=True, indent=4))
# save ocr.xml
try:
with open('ocr.xml', 'w', encoding='utf-8') as file:
ocr = poll_fetch(page_json['ocr'])
if ocr is None:
log.warn("Could not fetch ALTO file. Skipping.")
for file in os.listdir():
os.remove(file)
os.chdir(root)
os.removedirs(path)
continue
file.write(ocr.text)
log.debug("OCR file: ./" + path + "/ocr.xml")
except Exception as e:
# skip to the next page
log.warn("OCR file not found: {}".format(e))
os.remove('ocr.xml')
os.remove('metadata.json')
os.chdir(root)
os.removedirs(path)
continue
if '--annotate' in sys.argv:
try:
# save scan.jp2
data = requests.get(page_json['jp2'], stream=True)
if data.status_code == 200:
with open('scan.jp2', 'wb') as file:
for chunk in data:
file.write(chunk)
log.debug("Image scan: ./" + path + "/scan.jp2")
else:
log.warn("Failed to get image scan. Skipping.")
except Exception as e:
log.error("Could not get image: {}".format(e))
log.warn("Skipping annotation stage for this page.")
handle_page(root)
# log.info("Cleaning up...")
for file in os.listdir():
os.remove(file)
os.chdir(root)
os.removedirs(path)
log.info()
log.dedent()
log.dedent()
log.info("Done fetching this batch, onto next")
log.warn("Can't fetch batch list right now. Perhaps site is down?")
log.warn("Please try again later.")
sys.exit(1)
try:
while True:
log.info("Batch page: {}".format(start_page))
log.indent()
for batch_info in batches_page_json['batches']:
log.info('Batch: ' + batch_info['name'])
if (curr_batch < start_batch):
pass
# unit testing
#elif(curr_batch > start_batch):
# sys.exit(0)
else:
fetch_batch(batch_info['name'], start_issue)
curr_batch = curr_batch + 1
log.dedent()
try:
batches_page_json = poll_fetch(
batches_page_json['next']).json()
except KeyError:
log.info("Congratulations! You've reached the end!!!")
break
except KeyboardInterrupt:
print("")
log.level = 0
log.info("Stopping script.")