def get_page_images (self):
tf = mktempfile(".pnm")
cmd = ('"%s" --input-format jp2 --input "%s" --output-format pnm --output "%s"'
% (JASPER, self.doc, tf))
try:
status, output, tsignal = subproc(cmd)
if status == 0:
# success
img = Image.open(tf)
imagespath = self.images_path()
os.mkdir(imagespath)
if self.uses_png:
png_file_name = os.path.join(imagespath, "page00001.png")
img.save(png_file_name, "PNG")
else:
if (convert_image_to_tiff(tf, imagespath)):
note(3, "created tiff file in %s", imagespath)
else:
note("Can't convert %s. Output was %s.", self.doc, output)
note(4, "cmd was %s", cmd)
note(4, "tfile %s %s", tf, (os.path.exists(tf) and "exists") or "does not exist")
raise RuntimeError(output)
finally:
if os.path.exists(tf):
os.unlink(tf)
def get_images_for_page (page_index, wordboxes, dpi, images_dir):
pageimages = []
filepath = os.path.join(images_dir, "page%05d.png" % (page_index + 1))
if os.path.exists(filepath):
wordboxes_file = tempfile.mktemp()
try:
boxlist = []
if wordboxes:
# first, write out list of wordboxes, in Leptonica BOXA format
for i in range(len(wordboxes)):
box = boxes[i]
x, y, w, h = (int(box.left() * dpi / 72.0), int(box.top() * dpi / 72.0),
int(box.width() * dpi / 72.0), int(box.height() * dpi / 72.0))
if (w > 0) and (h > 0):
boxlist.append((x, y, w, h))
if len(boxlist) > 0:
fp = open(wordboxes_file, "wb")
fp.write("\nBoxa Version 2\nNumber of boxes = %d\n" % len(boxlist))
for i in range(len(boxlist)):
fp.write(" Box[%d]: " % i + "x = %d, y = %d, w = %d, h = %d\n" % boxlist[i])
fp.close()
# now, run the finder on the page image plus the list of wordboxes
debug_arg = (debug and "--debug") or " "
cmd = "%s %s %s %s %s" % (FINDIMAGES_PROGRAM, debug_arg, dpi, filepath, (boxlist and wordboxes_file) or "-")
note(4, "findimages cmd is <<%s>>", cmd)
status, output, tsignal = subproc(cmd)
if status == 0:
for line in [x.strip() for x in output.split('\n') if x.strip()]:
if not line.startswith("halftone "):
continue
pageimages.append((str(page_index) + " " + line.strip()).split())
else:
note(3, "findimages command <%s> returns bad status %s:\n%s\n" % (cmd, status, output))
finally:
# remove the temp file
if os.path.exists(wordboxes_file):
os.unlink(wordboxes_file)
# note("%d: wordboxes file is %s", page_index, wordboxes_file)
return pageimages
def process_tarred_folder (repo, id, tarfile, metadata):
# create a new folder, and populate it
dirname = tempfile.mktemp()
try:
os.mkdir(dirname)
os.chmod(dirname, 0700)
cmd = UNTAR_CMD % (dirname, TAR, tarfile)
note(2, "Untarring folder into temporary directory %s", dirname)
status, output, signal = subproc(cmd)
if status == 0:
note(2, "Successfully untarred folder into %s", dirname)
if metadata:
update_metadata(os.path.join(dirname, "metadata.txt"), metadata)
if (os.path.exists(os.path.join(dirname, "document.tiff")) or
os.path.isdir(os.path.join(dirname, "page-images"))):
return process_folder(repo, id, dirname, true)
else:
raise Error("invalid folder -- no page images file")
else:
raise Error("Problem untarring folder:\n%s" % output)
finally:
if os.path.exists(dirname):
shutil.rmtree(dirname)
def index_folder (folder, repo_index_dir):
update_configuration()
docs_dir, doc_id = os.path.split(folder)
SECTION_LOCK.acquire()
try:
try:
if LUCENE == 'jcc':
c = get_context(repo_index_dir)
c.index(folder, doc_id)
else:
indexingcmd = INDEXING_ADD_CMD % (JAVA, DEBUG_FLAGS, INDEXING_PROPERTIES, LUCENE_JAR, INDEXING_JAR, repo_index_dir, docs_dir, doc_id)
note(3, " indexing with %s", indexingcmd)
status, output, tsignal = subproc(indexingcmd)
except:
note(0, "Can't index folder %s:\n%s",
folder, ''.join(traceback.format_exception(*sys.exc_info())))
finally:
SECTION_LOCK.release()
if LUCENE != 'jcc':
note(3, " indexing output is <%s>", output)
if status != 0:
raise Error ("%s signals non-zero exit status %d attempting to index %s:\n%s" % (JAVA, status, folder, output))
def remove_from_index (repo_index_dir, doc_id):
update_configuration()
if LUCENE == 'jcc':
SECTION_LOCK.acquire()
try:
c = get_context(repo_index_dir)
c.remove(doc_id)
finally:
SECTION_LOCK.release()
else:
indexingcmd = INDEXING_REMOVE_CMD % (JAVA, DEBUG_FLAGS, "", LUCENE_JAR, INDEXING_JAR, repo_index_dir, doc_id)
note(3, " de-indexing with %s", indexingcmd)
SECTION_LOCK.acquire()
try:
status, output, tsignal = subproc(indexingcmd)
finally:
SECTION_LOCK.release()
note(3, " indexing output is <%s>", output)
if status != 0:
raise Error ("%s signals non-zero exit status %d attempting to remove %s:\n%s" % (JAVA, status, doc_id, output))
def do_thumbnails (dirpath, output_dir, **params):
note(2, " thumbnailing in %s...", dirpath)
tmpdir = tempfile.mktemp()
retval = params.get('returnvalue', false)
doc_metadata_path = os.path.join(dirpath, "metadata.txt")
try:
os.mkdir(tmpdir)
os.chmod(tmpdir, 0700)
try:
md = read_metadata(doc_metadata_path)
is_temporary_doc = md.get("temporary-contents")
if is_temporary_doc and (is_temporary_doc == "true"):
# temporary -- don't spend much time on this
create_temporary_icons (md, dirpath, output_dir, params)
retval = true
return
if os.path.exists(os.path.join(dirpath, "document.tiff")):
# contains one single-page TIFF file
tiffmaster = os.path.join(tmpdir, "master.tiff")
split_command = (TIFF_SPLIT_CMD
% (TIFFCP, os.path.join(dirpath, "document.tiff"), tiffmaster,
TIFFSPLIT, tiffmaster, os.path.join(tmpdir, "x")))
status, output, tsignal = subproc(split_command)
if status != 0:
raise Error ("'%s' signals non-zero exit status %d in %s => %s"
% (split_command, status, dirpath, tmpdir))
parts_dir = tmpdir
filecheck_fn = lambda fn: fn[0] == "x"
elif (os.path.exists(os.path.join(dirpath, "page-images")) and
os.path.isdir(os.path.join(dirpath, "page-images"))):
# contains directory full of PNG page images
parts_dir = os.path.join(dirpath, "page-images")
filecheck_fn = lambda fn: (fn.startswith('page') and fn.endswith('.png'))
else:
raise Error("No page images for document in %s" % dirpath)
tiff_parts = os.listdir(parts_dir)
if len(tiff_parts) < 1:
raise Error("No pages in split tiff file directory after split!")
# either a PNG-images or a TIFF split will sort properly in lexicographic order
tiff_parts.sort()
# see if there's a document icon legend and info about the DPI of the tiff file
legend = md.get('document-icon-legend')
tiff_dpi = int(md.get('images-dpi') or md.get('tiff-dpi') or params.get('images-dpi') or 0)
page_numbers_v = md.get('page-numbers')
page_numbers = (page_numbers_v and figure_page_numbers(page_numbers_v, dirpath))
first_page = int(md.get('first-page-number', 1))
skips = md.get('document-bbox-pages-to-skip', '')
if skips:
parts = string.split(skips, ':')
bbox_skips = []
for part in parts:
bbox_skips = bbox_skips + map(int, string.split(part, ','))
else:
bbox_skips = None
# figure bounding box for imaged page
page_count = 0
bbox = None
note(2, " calculating bounding box for large pages...")
dont_crop = md.get('dont-crop-big-thumbnails', false)
if AUTO_CROP_BIG_THUMBNAILS and not dont_crop:
do_bbox = true
else:
do_bbox = false
for tiff_part in tiff_parts:
if not filecheck_fn(tiff_part):
continue
if page_count == 0:
# find the width and height of the document
docwidth, docheight = figure_doc_size(os.path.join(parts_dir, tiff_part))
if not do_bbox:
bbox = (0, 0, docwidth, docheight)
if do_bbox:
bbox = figure_bbox (os.path.join(parts_dir, tiff_part), page_count, bbox, bbox_skips)
if (bbox and bbox[0] == 0) and (bbox[1] == 0) and (bbox[2] >= docwidth) and (bbox[3] >= docheight):
# don't bother, there's no area to crop already
do_bbox = false
page_count = page_count + 1
if page_count == 0:
raise Error("No pages in split tiff file directory after split!")
note(2, " final bbox is %s, page_count is %d", bbox, page_count)
if USE_VIRTUAL_INK:
note(2, " alpha channels will be added to large thumbnails...")
# now make the thumbnails
big_thumbnail_size = []
small_thumbnail_size = []
icon_size = []
page_index = 0
for tiff_part in tiff_parts:
if not filecheck_fn(tiff_part):
note(3, " skipping %s", tiff_part)
continue
tiff_path = os.path.join(parts_dir, tiff_part)
if page_numbers:
page_no_string = page_numbers.get(page_index)
else:
page_no_string = None
note (2, " page %d%s", page_index, (page_no_string and " (%s)" % page_no_string) or "")
try:
if not create_thumbnail(tiff_path, tiff_dpi, output_dir,
page_index, first_page, page_count, bbox, bbox_skips,
big_thumbnail_size, small_thumbnail_size, icon_size,
params.get('maxwidth'), params.get('maxheight'), params.get('maxscaling'),
params.get('thumbnail_strategy'), legend, page_no_string):
raise Error ("Can't create thumbnail for page %d in %s (of %s)" % (page_index, tiff_path, dirpath))
except Exception, x:
doc_id = os.path.split(dirpath)[1]
note("exception creating thumbnails for page %d of document %s:\n%s", page_index, doc_id,
string.join(traceback.format_exception(*sys.exc_info()), ""))
raise AbortDocumentIncorporation(doc_id, str(x))
if page_index == 0:
bt_width = big_thumbnail_size[0]
bt_height = big_thumbnail_size[1]
st_width = small_thumbnail_size[0]
st_height = small_thumbnail_size[1]
else:
bt_width = max(bt_width, big_thumbnail_size[0])
bt_height = max(bt_height, big_thumbnail_size[1])
st_width = max(st_width, small_thumbnail_size[0])
st_height = max(st_height, small_thumbnail_size[1])
st_scaling = (float(st_width)/float(docwidth) + float(st_height)/float(docheight)) / 2.0
page_index = page_index + 1
d = {"page-count" : str(page_count),
"tiff-width" : str(docwidth),
"images-width" : str(docwidth),
"images-size" : "%d,%d" % (docwidth, docheight),
"cropping-bounding-box" : "%d,%d;%d,%d" % (bbox),
"big-thumbnail-size" : "%s,%s" % (bt_width, bt_height),
"small-thumbnail-size" : "%s,%s" % (st_width, st_height),
"small-thumbnail-scaling" : "%f" % st_scaling,
"icon-size" : "%d,%d" % icon_size[0],
"images-height" : str(docheight),
"tiff-height" : str(docheight) }
translation, scaling = thumbnail_translation_and_scaling(dirpath, d, false, true)
d["big-thumbnail-translation-points"] = "%f,%f" % translation
d["big-thumbnail-scaling-factor"] = "%f,%f" % scaling
update_metadata(os.path.join(dirpath, "metadata.txt"), d)
finally:
shutil.rmtree(tmpdir)
# indicate successful completion
note(2, " finished.")
retval = true
def _add_internal (ostream, percent_done_fn, repo, response, params, content, wait):
# this can be called in several different ways.
# In general, you post a multipart/form-data body which
# contains a "contenttype" for the document, and either a "URL"
# for the content, or a "content" parameter containing the
# the actual content. If both "URL" and "content" are present,
# the URL is added as the "original-url" value for the metadata,
# and if the content is HTML, it's used as the "original.html"
# and the URL is used to pull ancillary content referenced in it.
content_type = params.get("contenttype")
url = params.get("URL")
noredir = params.get("no-redirect")
noredir = noredir and (noredir.lower() == "true")
uploadloc = url
docname = params.get("documentname")
tempf = None
suppress_duplicates = params.get("suppress-duplicates")
suppress_duplicates = suppress_duplicates and (suppress_duplicates.lower() == "true")
bury = params.get("bury")
bury = bury and (bury.lower() == "true")
verbosity = int(params.get("verbosity") or "0")
if content:
if wait and ostream:
_rewrite_job_output(ostream, '{ state: 0, msg: "Caching page..."}')
extension = CONTENT_TYPES.get(content_type)
if not extension:
if wait:
msg = "Don't know what to do with contenttype \"%s\"" % content_type
if ostream:
_rewrite_job_output(ostream, '{state: 1, msg: "' + urllib.quote(msg) + '"}')
else:
response.error(HTTPCodes.UNSUPPORTED_MEDIA_TYPE, msg)
return
# special case HTML/XHTML
if content and (content_type.lower() in ("text/html", "application/xhtml+xml")):
tempf = tempfile.mkdtemp()
uploadloc = os.path.join(tempf, "original.html")
# make sure that the folder for other parts exists, even if empty
os.mkdir(os.path.join(tempf, "original_files"))
# remove our bookmarklet, if present
content = _BOOKMARKLET_PATTERN.sub('', content)
content = _ADD_FORM_PATTERN.sub('', content)
c = _OurCacher(url, filename=uploadloc, bits=content, content_type=content_type)
# make sure that the folder for other parts exists, even if empty
other_parts = os.path.join(tempf, "original_files")
if not os.path.exists(other_parts):
os.mkdir(other_parts)
# special case 3x5 cards
elif (docname and (content_type.lower() == "text/plain") and os.path.splitext(docname)[1] == ".3x5"):
fd, tempf = tempfile.mkstemp(".3x5")
fp = os.fdopen(fd, "wb")
fp.write(content)
fp.close()
uploadloc = tempf
else:
fd, tempf = tempfile.mkstemp("." + extension)
fp = os.fdopen(fd, "wb")
fp.write(content)
fp.close()
uploadloc = tempf
if suppress_duplicates:
hash = calculate_originals_fingerprint(tempf)
results = repo.do_query("sha-hash:"+hash)
if results:
# it's a duplicate
doc = results[0][1]
if os.path.isdir(tempf):
shutil.rmtree(tempf)
elif os.path.exists(tempf):
os.remove(tempf)
if ostream:
_rewrite_job_output(ostream, '{ state: 2, doc_id: "' + doc.id + '"}')
elif noredir:
response.reply(doc.id, "text/plain")
else:
response.redirect("/action/basic/dv_show?doc_id=%s" % doc.id)
return
try:
try:
# get a cookie for authentication
cookie = repo.new_cookie(url or content[:min(100, len(content))])
cookie_str = '%s=%s; path=/; Secure' % (cookie.name(), cookie.value())
os.environ["UPLIB_COOKIE"] = cookie_str
doctitle = params.get("md-title")
docauthors = params.get("md-authors")
docdate = params.get("md-date")
doccats = params.get("md-categories")
metadata = params.get("metadata")
if metadata:
mdtmpfile = tempfile.mktemp()
open(mdtmpfile, "w").write(metadata)
# check to see if we're replacing an existing document
md2 = read_metadata(StringIO.StringIO(metadata))
existing_doc_id = md2.get("replacement-contents-for")
if existing_doc_id and not repo.valid_doc_id(existing_doc_id):
raise ValueError("Invalid doc ID %s specified for replacement" % existing_doc_id)
else:
mdtmpfile = None
existing_doc_id = None
# now form the command
scheme = ((repo.get_param("use-http", "false").lower() == "true") or _use_http) and "http" or "https"
cmd = '%s --verbosity=%s --repository=%s://127.0.0.1:%s ' % (_uplib_add_document, verbosity, scheme, repo.port())
if doctitle:
cmd += ' --title=%s' % pipes.quote(doctitle)
if docauthors:
cmd += ' --authors=%s' % pipes.quote(docauthors)
if docdate:
cmd += ' --date="%s"' % docdate
if doccats:
cmd += ' --categories=%s' % pipes.quote(doccats)
if mdtmpfile:
cmd += ' --metadata="%s"' % mdtmpfile
cmd += ' "%s"' % uploadloc
if ostream:
_rewrite_job_output(ostream, '{state: 0, msg: "' + urllib.quote(cmd) + '"}')
# and invoke the command
status, output, tsignal = subproc(cmd)
note(4, "cmd is %s, status is %s, output is %s", repr(cmd), status, repr(output.strip()))
if mdtmpfile:
os.unlink(mdtmpfile)
if status == 0:
# success; output should be doc-id
doc_id = existing_doc_id or output.strip().split()[-1]
note(4, "output is '%s'; doc_id for new doc is %s", output.strip(), doc_id)
if wait and ostream:
_rewrite_job_output(ostream, '{ state: 1, doc_id: "' + doc_id + '", msg: "' + urllib.quote(output) + '"}')
# wait for it to come on-line
if percent_done_fn:
percent_done_fn(40) # estimate 40% of work done on client side
while not repo.valid_doc_id(doc_id):
if ostream:
pending = repo.list_pending(full=True)
s = _first(pending, lambda x: x['id'] == doc_id)
if not s:
break
dstatus = s['status']
if dstatus == 'error':
msg = 'server-side error incorporating document'
_rewrite_job_output(ostream, '{ state: 3, doc_id: "' + doc_id
+ '", msg: "' + urllib.quote(s['error']) + '"}')
break
if dstatus == 'unpacking':
msg = 'starting ripper process...'
elif dstatus == 'ripping':
msg = "ripping with ripper '" + s['ripper'] + "'..."
elif dstatus == 'moving':
msg = 'adding to registered document set...'
_rewrite_job_output(ostream, '{ state: 1, doc_id: "' + doc_id
+ '", msg: "' + urllib.quote(msg) + '"}')
time.sleep(1.0)
if percent_done_fn:
percent_done_fn(100) # finished
if repo.valid_doc_id(doc_id):
if bury:
# wait up to 100 seconds for it to show up in history list
# after that, wait another second, then bury it
counter = 100
while counter > 0:
h = [x.id for x in repo.history()]
if doc_id in h:
break
counter -= 1
time.sleep(1)
time.sleep(1)
repo.touch_doc(doc_id, bury=True, notify=False)
note(3, "buried %s", doc_id)
if wait:
if ostream:
_rewrite_job_output(ostream, '{ state: 2, doc_id: "' + doc_id + '"}')
elif noredir:
response.reply(doc_id, "text/plain")
else:
response.redirect("/action/basic/dv_show?doc_id=%s" % doc_id)
else:
note("cmd <<%s>> failed with status %s:\n%s", cmd, status, output)
if wait:
if ostream:
_rewrite_job_output(ostream, '{ state: 3, msg: "' + urllib.quote('Error processing the document:\n' + output) + '"}')
else:
response.error(HTTPCodes.INTERNAL_SERVER_ERROR, "<pre>" + htmlescape(output) + "</pre>")
except:
e = ''.join(traceback.format_exception(*sys.exc_info()))
if wait:
note(3, "Exception processing uplib-add-document request:\n%s", htmlescape(e))
if ostream:
_rewrite_job_output(ostream, '{state: 3, msg: "' + urllib.quote("Exception processing uplib-add-document request:\n" + e) + '"}')
else:
response.error(HTTPCodes.INTERNAL_SERVER_ERROR,
"Exception processing uplib-add-document request:\n<pre>" +
htmlescape(e) + "\n</pre>")
else:
note("Exception processing uplib-add-document request:\n%s", e)
finally:
if tempf and os.path.isfile(tempf):
os.unlink(tempf)
elif tempf and os.path.isdir(tempf):
shutil.rmtree(tempf)
def flesh_out_folder(id, tmpfilename, metadata, repo, unpack_fn, counter):
try:
try:
# note(3, "CODETIMER_ON is %s", CODETIMER_ON)
# if CODETIMER_ON:
# code_timer.Init()
# code_timer.CreateTable("uplib")
# code_timer.CodeTimerOn()
# code_timer.StartInt("newFolder$unpack", "uplib")
# else:
# code_timer.CodeTimerOff()
if unpack_fn and tmpfilename and os.path.exists(tmpfilename):
unpack_fn(repo, id, tmpfilename, metadata)
# if CODETIMER_ON:
# code_timer.StopInt("newFolder$unpack", "uplib")
folderpath = repo.pending_location(id)
try:
note("unpacked new folder in %s", folderpath)
if not sys.platform.lower().startswith("win"):
s, o, t = subproc("ls -Rl %s" % folderpath)
note("%s\n" % o)
fp = open(os.path.join(folderpath, "UNPACKED"), 'w')
fp.flush()
fp.close()
# as of this point, we can restart the inclusion of the document
md = read_metadata(os.path.join(folderpath, "metadata.txt"))
replacement_id = md.get("replacement-contents-for")
if replacement_id:
if repo.valid_doc_id(replacement_id):
# contents to replace another document
md["replacement-contents-for"] = ""
update_metadata(os.path.join(folderpath, "metadata.txt"), md)
note(2, "replacing contents of %s with this data...", replacement_id)
existing_document = repo.get_document(replacement_id)
new_folder = existing_document.folder()
process_folder(repo, replacement_id, folderpath, false, new_folder)
_run_rippers(new_folder, repo, replacement_id)
existing_document.recache()
repo.touch_doc(existing_document)
raise AbortDocumentIncorporation(id, "replacement for existing document %s" % replacement_id)
else:
raise AbortDocumentIncorporation(id, "replacement for non-existent document %s" % replacement_id)
_finish_inclusion (repo, folderpath, id)
# if CODETIMER_ON:
# noteOut = StringIO.StringIO()
# noteOut.write("\nCode Timer statistics (what took time, in milliseconds):\n")
# code_timer.PrintTable(noteOut, "uplib")
# noteOut.write("\n")
# noteOutString = noteOut.getvalue()
# note(3, noteOutString)
except:
type, value, tb = sys.exc_info()
note("%s", ''.join(traceback.format_exception(type, value, tb)))
note_error(folderpath, (type, value, tb))
raise value, None, tb
except AbortDocumentIncorporation, x:
# ripper signalled to stop adopting this document, for good
note(2, "AbortDocumentIncorporation exception on %s: %s", x.id, x.message)
if (x.id == id):
shutil.rmtree(folderpath)
remove_from_index(repo.index_path(), id)
except:
type, value, tb = sys.exc_info()
note("Exception processing new folder:\n%s", ''.join(traceback.format_exception(type, value, tb)))
def index_folders (docs_dir, doc_ids, repo_index_dir):
update_configuration()
if not doc_ids:
return
if LUCENE == 'jcc':
c = get_context(repo_index_dir)
SECTION_LOCK.acquire()
try:
for id in doc_ids:
folderpath = os.path.join(docs_dir, id)
if os.path.isdir(folderpath):
lock_folder(folderpath)
try:
try:
c.index(folderpath, id, False)
except:
note(0, "Can't index folder %s:\n%s",
folderpath, ''.join(traceback.format_exception(*sys.exc_info())))
finally:
unlock_folder(folderpath)
c.reopen()
finally:
SECTION_LOCK.release()
return
else:
# invoke Java to do indexing
if len(doc_ids) > 6:
fname = tempfile.mktemp()
fp = open(fname, "w")
fp.write(string.join(doc_ids, '\n'))
fp.close()
indexingcmd = INDEXING_BATCHADD_CMD % (JAVA, DEBUG_FLAGS, INDEXING_PROPERTIES, LUCENE_JAR, INDEXING_JAR, repo_index_dir, docs_dir, fname)
note(3, " indexing with %s", indexingcmd)
SECTION_LOCK.acquire()
try:
status, output, tsignal = subproc(indexingcmd)
finally:
SECTION_LOCK.release()
os.unlink(fname)
note(3, " indexing output is <%s>", output)
if status != 0:
raise Error ("%s signals non-zero exit status %d attempting to index %s:\n%s" % (JAVA, status, doc_ids, output))
else:
folders = string.join(doc_ids, ' ')
indexingcmd = INDEXING_ADD_CMD % (JAVA, DEBUG_FLAGS, INDEXING_PROPERTIES, LUCENE_JAR, INDEXING_JAR, repo_index_dir, docs_dir, folders)
note(3, " indexing with %s", indexingcmd)
SECTION_LOCK.acquire()
try:
status, output, tsignal = subproc(indexingcmd)
finally:
SECTION_LOCK.release()
note(3, " indexing output is <%s>", output)
if status != 0:
raise Error ("%s signals non-zero exit status %d attempting to index %s:\n%s" % (JAVA, status, doc_ids, output))
def manipulate_server_internal (repo, params, response=None, ipaddr=None, lgr=None):
# regular UpLib action
conf = params.get("configurator")
if not conf:
conf = configurator()
imap_ssl_port = conf.get_int("imap-server-ssl-port", -1)
imap_localhost_port = conf.get_int("imap-server-localhost-port", 8143)
stunnel = conf.get("stunnel")
expunge_deletes_docs = conf.get_bool("imap-expunge-deletes-documents", False)
global CHECKPOINT_PERIOD
CHECKPOINT_PERIOD = conf.get_int("imap-server-checkpoint-interval", 600)
allow_anonymous_readers = ((not repo.has_password) and
conf.get_bool("imap-server-allow-anonymous-readers", True))
use_for_email = conf.get_bool("imap-server-use-for-email", False)
imap_dir = os.path.join(repo.overhead_folder(), "imap")
if not os.path.isdir(imap_dir):
os.mkdir(imap_dir)
stunnel_pid_filepath = os.path.join(imap_dir, "stunnel.pid")
if os.path.exists(stunnel_pid_filepath):
stunnel_pid = int(open(stunnel_pid_filepath, 'r').read().strip())
else:
stunnel_pid = None
# we cache the reference to the existing server in another
# module so that we can reload this one with impunity
current_server = emailParser.__dict__.get("IMAP_SERVER")
note("current server is %s", current_server)
action = params.get('action')
newcontext = params.get('newcontext', False)
if response:
fp = response.open()
else:
fp = StringIO()
fp.write('<body bgcolor="%s">\n' % STANDARD_BACKGROUND_COLOR)
if current_server:
s = current_server.status()
m = s.more()
while m:
fp.write(m)
m = s.more()
fp.write('\n<hr>\n')
else:
fp.write('<h2>UpLib IMAP Server control panel</h2>\n')
current_context = None
if current_server and ((action == 'Stop') or (action == 'Restart')):
if stunnel_pid:
try:
os.kill(stunnel_pid, signal.SIGKILL)
time.sleep(4)
except:
pass
stunnel_pid = None
current_context = current_server.mailcontext
current_server.close()
current_server = None
del emailParser.__dict__["IMAP_SERVER"]
fp.write("<p>Closed current server.\n")
if os.path.exists(stunnel_pid_filepath):
os.unlink(stunnel_pid_filepath)
if (action == 'Start') or (action == 'Restart'):
cert_filepath = os.path.join(repo.overhead_folder(), repo.certfilename())
try:
port = params.get("port")
if port:
port = int(port)
else:
port = imap_localhost_port
if stunnel and ((not ssl) or (imap_ssl_port > 0)):
# start stunnel
stunnel_conf_filepath = os.path.join(imap_dir, "stunnel.conf")
f = open(stunnel_conf_filepath, 'w')
f.write("debug = 7\n\ncert = %s\noutput = %s\npid = %s\n\n[imapuplib]\naccept = %s\nconnect = 127.0.0.1:%s\n" %
(cert_filepath, os.path.join(imap_dir, "stunnel.log"), stunnel_pid_filepath,
str(imap_ssl_port), str(port)))
f.close()
status, tsignal, output = subproc("%s %s" % (stunnel, stunnel_conf_filepath))
note("status from '%s %s' (on %s) is %s, output is <%s>", stunnel, stunnel_conf_filepath, imap_ssl_port, status, output)
if status != 0:
raise RuntimeError("Can't start stunnel with '%s %s'; status is %s, output is %s" % (stunnel, stunnel_conf_filepath, status, output))
stunnel_pid = int(open(stunnel_pid_filepath, 'r').read().strip())
note("stunnel_pid is %s", stunnel_pid)
else:
stunnel_pid = None
if newcontext or (not current_context):
current_context = uplib_mailcontext(repo,
expunge_deletes_docs=expunge_deletes_docs,
allow_readers=allow_anonymous_readers,
use_for_email=use_for_email,
ip=get_fqdn(),
server_certificate_file=cert_filepath)
if current_context.inbox:
current_context.inbox.rescan()
if stunnel_pid is not None:
ipaddr = '127.0.0.1'
else:
ipaddr = '0.0.0.0'
if not lgr:
lgr = logger.rotating_file_logger (os.path.join(imap_dir, "imap.log"), "weekly", None, True)
lgr = logger.unresolving_logger(lgr)
imaps = imap_server (current_context, ipaddr, port, logger=lgr, stunnel_pid=stunnel_pid)
emailParser.__dict__["IMAP_SERVER"] = imaps
current_server = imaps
hooked = emailParser.__dict__.get("IMAP_SERVER_SHUTDOWN_HOOK")
if not hooked:
repo.add_shutdown_hook(lambda x=repo: shutdown_server(x))
emailParser.__dict__["IMAP_SERVER_SHUTDOWN_HOOK"] = True
if stunnel_pid:
fp.write("<p>Started new IMAP4 server for %s on ports %s/%s."
% (repr(repo), str(imap_ssl_port), str(port)))
else:
fp.write("<p>Started new IMAP4 server for %s on port %s."
% (repr(repo), str(port)))
if current_context.inbox:
fp.write("<p>Inbox: %d messages, %d recent, %d unseen."
% (len(current_context.inbox.msgs),
len(current_context.inbox.recent()),
current_context.inbox.min_unseen()))
except:
type, value, tb = sys.exc_info()
s = string.join(traceback.format_exception(type, value, tb))
note("Can't establish IMAP server: exception: " + s)
fp.write(s)
fp.write('<form method=GET action="/action/IMAPServer/manipulate_server">\n')
fp.write('<input type=submit name=action value="Start" %s>\n' % ((current_server and "disabled") or ""))
fp.write('<input type=submit name=action value="Stop" %s>\n' % (((current_server == None) and "disabled") or ""))
fp.write('<input type=submit name=action value="Restart" %s>\n' % (((current_server == None) and "disabled") or ""))
fp.write('<input type=checkbox name="newcontext" %s> Use fresh mail context\n' % (newcontext and "checked") or "")
fp.write('</form>\n')
fp.write('</body>\n')
def rip (self, location, doc_id):
global CITATION_PARSER, HEADER_PARSER
omd = self.get_folder_metadata(location)
# CiteSeer really only works on traditional publications, so let's stay
# with PDF and Word docs
mimetype = omd.get("apparent-mime-type")
if mimetype not in TRADITIONAL_PAPER_FORMATS:
return
text, language = self.get_folder_text(location)
if not text:
# no text to look at
return
m = REFERENCES_PATTERN.search(text)
if not m:
# no REFERENCES_PATTERN in text
return
# just a note if we're re-ripping something
if self.repository().valid_doc_id(doc_id):
note(3, "%s is a technical report", self.repository().get_document(doc_id))
cp = self.__citation_parser or CITATION_PARSER
if cp:
status, output, tsig = subproc('%s "%s"' % (cp, self.folder_text_path(location)))
if status == 0:
parsed = BeautifulStoneSoup(output.strip())
citations = parsed.findAll("citation")
note(3, "found %d citations", len(citations))
fp = open(os.path.join(location, "citeseerx-citations.xml"), "w")
fp.write(output.strip())
hp = self.__header_parser or HEADER_PARSER
if hp:
tfile = tempfile.mktemp()
fp = codecs.open(tfile, "w", "UTF-8")
fp.write(text)
fp.close()
try:
status, output, tsig = subproc('%s "%s"' % (hp, tfile))
if status == 0:
md = dict()
parsed = BeautifulStoneSoup(output.strip())
title = parsed.find("title")
if title:
if title.string:
md['citeseer-title'] = title.string
else:
note(3, "Non-string title found: %s", title)
authors = set()
for author in parsed.findAll("author"):
n = author.find("name")
if n:
authors.add(n.string)
else:
authors.add(author.string)
if authors:
md['citeseer-authors'] = " and ".join(list(authors))
abstract = parsed.find("abstract")
if abstract:
if abstract.string:
md['citeseer-abstract'] = abstract.string
else:
note(3, "Non-string abstract found: %s", abstract)
note(3, "citeseer metadata is %s", pprint.pformat(md))
if "citeseer-title" in md:
# use CiteSeer data to fix up document metadata, if necessary
if ((not omd.get("title")) or
(omd.get("title-is-original-filepath", "false").lower() == "true")):
md['title'] = md.get("citeseer-title")
md['title-is-original-filepath'] = None
md['title-is-citeseer-extracted'] = "true"
if ("citeseer-authors" in md) and (not omd.get("authors")):
md['authors'] = md.get("citeseer-authors")
if ("citeseer-abstract" in md) and (not md.get("abstract")):
abs = md.get("citeseer-abstract")
prefix = ABSTRACT_PREFIX.match(abs)
if prefix:
realstart = prefix.end("prefix")
note(3, "trimming abstract prefix of %s", repr(abs[:realstart]))
abs = abs[realstart:]
md['abstract'] = abs
note(3, "updated missing metadata with CiteSeer versions")
self.update_folder_metadata(location, md)
finally:
if os.path.exists(tfile):
os.unlink(tfile)
