def get_info(pid, words=None, extra_previews=True):
def b64img(img):
data = io.BytesIO()
img.save(data, format='JPEG', quality=85)
return base64.b64encode(data.getvalue()).decode()
mh = MediaHaven()
alto = mh.get_alto(pid)
result = dict(
pid=pid,
words=len(words) if words is not None else 0,
alto=alto.search_words(words),
alto_link=alto.url
)
result['ocr_text'] = alto.text
with mh.get_preview(pid) as im:
result['previewImageUrl'] = im.meta['previewImagePath']
result['meta'] = im.meta
if result['words'] > 0 and extra_previews:
result['preview_full'] = b64img(im.highlight_words(words, crop=False))
result['preview'] = b64img(im.highlight_words(words))
result['props'] = im.meta['mdProperties']
return result
class Importer:
def __init__(self):
self._solr = Solr(Config(section='solr')['url'])
self._mh = MediaHaven(buffer_size=100)
def add(self, item):
self._solr.add([item])
def process(self, item):
if item is None:
raise Exception("Invalid item passed (None)")
if type(item) is not str:
pid = item['externalId']
else:
pid = item
item = self._mh.one('+(externalId:%s)' % pid)
if not pid:
raise "No pid for item %s" % (item, )
language = ''
try:
language = item['mdProperties']['language'][0].lower()
except Exception as e:
logger.warning('no language found for %s', pid)
logger.exception(e)
alto = self._mh.get_alto(item)
if not alto:
logger.debug("no alto for pid '%s' " % (pid, ))
text = ''
else:
text = Conversions.normalize(alto.text)
self.add(dict(id=pid, text=text, language=language))