def process_file(id_source):
from processing.tex_template import tex_header
from pyPdf import PdfFileReader
from numpy import array
logging.info("begin %s" % (id_source, ))
OUTPUT = []
S = {}
locations, comments = annotations.getPublicCommentsByFile(id_source)
repfile = "%s/%s/%s" % (settings.HTTPD_MEDIA,settings.REPOSITORY_DIR, id_source)
if not os.path.exists(repfile):
logging.warning("%s not found. Skipping..." % (repfile, ))
return
srcfile = "/tmp/orig_%s.pdf" % (id_source, )
if not os.path.exists(srcfile):
os.symlink(repfile, srcfile)
pdf = PdfFileReader(file(srcfile, "rb"))
if pdf.isEncrypted and pdf.decrypt("")==0:
print "PDF file encrypted with non-empty password: %s" % (srcfile,)
return False
trim_box = pdf.pages[0].trimBox # Sacha's coordinate system now uses this box
crop_box = pdf.pages[0].cropBox # ConTeXt's page inclusion uses this box
fudge = (int(trim_box[2])-int(trim_box[0]))/612.0 # for the assumption of 612bp width
bp_per_pixel = 72.0/150 * fudge
roots = {}
children_of = {}
for k in comments:
node = int(k)
parent = comments[k]['id_parent']
if parent:
if parent not in children_of:
children_of[parent] = []
children_of[parent].append(node)
else:
loc_id = comments[node]['ID_location']
loc = locations[loc_id]
if loc['page'] != 0:
loc['center_x'] = loc['left'] + loc['w']/2.0
loc['center_y'] = loc['top'] + loc['h']/2.0
else:
loc['center_x'] = None
loc['center_y'] = None
roots[node] = loc
def oneline(s):
return s.replace('\n', ' ')
def texify(s):
s = s.strip()
patterns = [(r'\\', r'\\\\'),
(r'%', r'\%'), (r'\$', r'\$'), ('_', r'\_'), (r'\&', r'\&'),
(r'\^', r'\^\\null{}'), (r'#', r'\#'), (r'\|', r'$|$')]
for p in patterns:
s = re.sub(p[0], p[1], s)
return s
def rect2array(rect):
return array(rect.lowerLeft+rect.upperRight, dtype=float)
def rectangle_height(rect):
return rect.upperRight[1]-rect.lowerLeft[1]
S["last_page"] = -1
def print_child(n, levels=0):
body = comments[n]['body']
loc_id = comments[n]['ID_location']
location = locations[loc_id]
page = int(location['page'])
if levels == 0 and page > S["last_page"]:
OUTPUT.append('\n%% Comments on page %d of %s [%s]' % (page,
"myfile",
os.path.basename(srcfile)))
if page == 0:
sectitle = 'Global comments'
else:
sectitle = 'Comments on page %d' % page
OUTPUT.append(r'\title{%s} \def\whatpage{%d}' % (sectitle, page))
S["last_page"] = page
if comments[n]['admin'] == 1:
me = 1
else:
me = 0
msg = '\n'+r'\comment{note-%s}{%d}{%s}{%d}{%d}' % (n, levels, texify(body), me, int(n))
OUTPUT.append(unicode(msg).encode("ascii", "ignore"))
if levels == 0 and page != 0: # a root comment not on page 0 needs callout
root = roots[n]
# Sacha's coords are from top left corner, relative to TrimBox
# but in pixels (not postscript points).
# evaluate comment_box_px, with this coord system, as [llx lly w h]
comment_box_px = array([root['left'],
root['top']+root['h'],
root['w'],
root['h']])
comment_box_bp = comment_box_px * bp_per_pixel
# convert y coordinate to use bottom edge of trim_box as y=0
comment_box_bp[1] = int(rectangle_height(trim_box))-int(comment_box_bp[1])
# convert to coordinates relative to CropBox
comment_box_bp[0:2] += (rect2array(trim_box)-rect2array(crop_box))[0:2]
OUTPUT.append('\setpospxywhd{note-%d-dest}{1}' % n)
OUTPUT.append('{%fbp}{%fbp}{%fbp}{%fbp}{0pt}' % tuple(comment_box_bp))
OUTPUT.append('''\startpositionoverlay{callouts}
\setMPpositiongraphic{note-%d}{callout}{to=note-%d-dest}
\stoppositionoverlay''' % (n, n))
if n in children_of:
for k in sorted(children_of[n]):
print_child(k, levels+1)
def cmp(a,b):
if roots[a]['page'] == 0 and roots[b]['page'] == 0:
return a-b # order by comment id
for key in ['page', 'center_y', 'center_x']:
if roots[a][key] != roots[b][key]:
return int(1e6*roots[a][key] - 1e6*roots[b][key])
return 0
tex_params = {'crop_wd': crop_box[2]-crop_box[0],
'crop_ht': crop_box[3]-crop_box[1],
'srcfile': srcfile
}
OUTPUT.append( tex_header % tex_params)
OUTPUT.append( '\n\\starttext')
for root_id in sorted(roots, cmp):
print_child(root_id, 0)
OUTPUT.append( '\n\\stoptext')
texfile = "/tmp/%s.tex" % (id_source, )
f = open(texfile, "w")
f.write("\n".join(OUTPUT))
f.close()
cmd = "(cd /tmp; texexec --timeout=120 %s; mv %s.pdf %s/%s/%s)" % (texfile, id_source, settings.HTTPD_MEDIA, settings.ANNOTATED_DIR, id_source)
os.system(cmd)
logging.info("end %s" % (id_source, ))
def process_file(id_source):
from pyPdf import PdfFileReader
from numpy import array
S = {}
locations, comments = annotations.getPublicCommentsByFile(id_source)
srcfile = "tmp/%s.pdf" % (id_source)
pdf = PdfFileReader(file(srcfile, "rb"))
if pdf.isEncrypted and pdf.decrypt("") == 0:
print "PDF file encrypted with non-empty password: %s" % (srcfile, )
return False
trim_box = pdf.pages[
0].trimBox # Sacha's coordinate system now uses this box
crop_box = pdf.pages[0].cropBox # ConTeXt's page inclusion uses this box
fudge = (int(trim_box[2]) -
int(trim_box[0])) / 612.0 # for the assumption of 612bp width
bp_per_pixel = 72.0 / 150 * fudge
roots = {}
children_of = {}
comments_res = {}
page_comment = {}
for k in comments:
node = int(k)
parent = comments[k]['id_parent']
if parent:
if parent not in children_of:
children_of[parent] = []
children_of[parent].append(node)
else:
loc_id = comments[node]['ID_location']
loc = locations[loc_id]
if loc['page'] != 0:
loc['center_x'] = loc['left'] + loc['w'] / 2.0
loc['center_y'] = loc['top'] + loc['h'] / 2.0
else:
loc['center_x'] = None
loc['center_y'] = None
roots[node] = loc
def oneline(s):
return s.replace('\n', ' ')
def texify(s):
s = s.strip()
patterns = [(r'\\', r'\\\\'), (r'%', r'\%'), (r'\$', r'\$'),
('_', r'\_'), (r'\&', r'\&'), (r'\^', r'\^\\null{}'),
(r'#', r'\#'), (r'\|', r'$|$')]
for p in patterns:
s = re.sub(p[0], p[1], s)
return s
def rect2array(rect):
return array(rect.lowerLeft + rect.upperRight, dtype=float)
def rectangle_height(rect):
return rect.upperRight[1] - rect.lowerLeft[1]
S["last_page"] = -1
def print_child(n, levels=0):
loc_id = comments[n]['ID_location']
location = locations[loc_id]
page = int(location['page'])
if levels == 0 and page > S["last_page"]:
S["last_page"] = page
if levels == 0 and page != 0: # a root comment not on page 0 needs callout
root = roots[n]
# Sacha's coords are from top left corner, relative to TrimBox
# but in pixels (not postscript points).
# evaluate comment_box_px, with this coord system, as [llx lly wwc h]
comment_box_px = array(
[root['left'], root['top'] + root['h'], root['w'], root['h']])
comment_box_bp = comment_box_px * bp_per_pixel
# convert y coordinate to use bottom edge of trim_box as y=0
comment_box_bp[1] = int(rectangle_height(trim_box)) - int(
comment_box_bp[1])
# convert to coordinates relative to CropBox
comment_box_bp[0:2] += (rect2array(trim_box) -
rect2array(crop_box))[0:2]
comments_res[n] = {}
comments_res[n]['location_ID'] = loc_id
comments_res[n]['source_ID'] = id_source
comments_res[n]['ensemble_ID'] = locations[loc_id]['id_ensemble']
comments_res[n]['x'] = comment_box_bp[0]
comments_res[n]['y'] = comment_box_bp[1]
comments_res[n]['w'] = comment_box_bp[2]
comments_res[n]['h'] = comment_box_bp[3]
comments_res[n]['page'] = page
comments_res[n]['parent'] = -1
strPage = str(page)
if strPage in page_comment:
tmp = page_comment[strPage]
tmp.append(n)
page_comment[strPage] = tmp
else:
page_comment[strPage] = [n]
elif levels != 0 and page != 0:
parent = comments[n]['id_parent']
comments_res[n] = {}
comments_res[n]['location_ID'] = loc_id
comments_res[n]['source_ID'] = id_source
comments_res[n]['ensemble_ID'] = locations[loc_id]['id_ensemble']
comments_res[n]['x'] = comments_res[parent]['x']
comments_res[n]['y'] = comments_res[parent]['y']
comments_res[n]['w'] = comments_res[parent]['w']
comments_res[n]['h'] = comments_res[parent]['h']
comments_res[n]['page'] = page
comments_res[n]['parent'] = parent
strPage = str(page)
if strPage in page_comment:
tmp = page_comment[strPage]
tmp.append(n)
page_comment[strPage] = tmp
else:
page_comment[strPage] = [n]
if n in children_of:
for k in sorted(children_of[n]):
print_child(k, levels + 1)
def cmp(a, b):
if roots[a]['page'] == 0 and roots[b]['page'] == 0:
return a - b # order by comment id
for key in ['page', 'center_y', 'center_x']:
if roots[a][key] != roots[b][key]:
return int(1e6 * roots[a][key] - 1e6 * roots[b][key])
return 0
for root_id in sorted(roots, cmp):
print_child(root_id, 0)
return comments_res, page_comment