Spaces:
Running
on
T4
Running
on
T4
# Licensed under the Apache License, Version 2.0 (the "License"); | |
# you may not use this file except in compliance with the License. | |
# You may obtain a copy of the License at | |
# | |
# http://www.apache.org/licenses/LICENSE-2.0 | |
# | |
# Unless required by applicable law or agreed to in writing, software | |
# distributed under the License is distributed on an "AS IS" BASIS, | |
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
# See the License for the specific language governing permissions and | |
# limitations under the License. | |
# | |
import copy | |
import re | |
from io import BytesIO | |
from PIL import Image | |
from rag.nlp import tokenize, is_english | |
from rag.nlp import rag_tokenizer | |
from ppt_parser import PptParser | |
class Ppt(PptParser): | |
def __call__(self, fnm, from_page, to_page, callback=None): | |
txts = super().__call__(fnm, from_page, to_page) | |
callback(0.5, "Text extraction finished.") | |
import aspose.slides as slides | |
import aspose.pydrawing as drawing | |
imgs = [] | |
with slides.Presentation(BytesIO(fnm)) as presentation: | |
for i, slide in enumerate(presentation.slides[from_page: to_page]): | |
buffered = BytesIO() | |
slide.get_thumbnail( | |
0.5, 0.5).save( | |
buffered, drawing.imaging.ImageFormat.jpeg) | |
imgs.append(Image.open(buffered)) | |
assert len(imgs) == len( | |
txts), "Slides text and image do not match: {} vs. {}".format(len(imgs), len(txts)) | |
callback(0.9, "Image extraction finished") | |
self.is_english = is_english(txts) | |
return [(txts[i], imgs[i]) for i in range(len(txts))] | |
def chunk(filename, binary=None, from_page=0, to_page=100000, | |
lang="English", callback=None, **kwargs): | |
""" | |
The supported file formats are pptx. | |
Every page will be treated as a chunk. And the thumbnail of every page will be stored. | |
PPT file will be parsed by using this method automatically, setting-up for every PPT file is not necessary. | |
""" | |
eng = lang.lower() == "english" | |
doc = { | |
"docnm_kwd": filename, | |
"title_tks": rag_tokenizer.tokenize(re.sub(r"\.[a-zA-Z]+$", "", filename)) | |
} | |
doc["title_sm_tks"] = rag_tokenizer.fine_grained_tokenize(doc["title_tks"]) | |
res = [] | |
ppt_parser = Ppt() | |
for pn, (txt, img) in enumerate( | |
ppt_parser(filename if not binary else binary, from_page, 1000000, callback) | |
): | |
d = copy.deepcopy(doc) | |
pn += from_page | |
d["image"] = img | |
d["page_num_int"] = [pn + 1] | |
d["top_int"] = [0] | |
d["position_int"] = [(pn + 1, 0, img.size[0], 0, img.size[1])] | |
tokenize(d, txt, eng) | |
res.append(d) | |
return res | |
raise NotImplementedError( | |
"file type not supported yet(pptx, pdf supported)") |