Python crawls all PDF documents in a single page

GitHub blog address, updates here may not be very timely.

1. Background

Recently found that the algorithm and data structure dropped a lot (in fact, the university is not how to learn, 囧rz), considering the recent project structure more and more complex, using it to practice ideas, it is intended to review the data structure and algorithm. Combine the recent learning of English and then simply use English. Then select a reference book "Data Structures and algorithms in Java".
At first, it was quite a struggle, slow down. As a result of the habit of ripping books appendix, so went to the book attached to the official website to see, found http://ww0.java4.datastructures.net/handouts/inside the PDF document is actually good, illustrated, as the understanding is a good material, Decisive to download AH. However, the results found that, a good many, this one to save as really deadly, think or what way to download it.

2. Implement

Consider all the languages that you have learned at the moment, which can be used to achieve and arrange the level:

1. Java/android familiar with
2. C # familiarity
3. Python Understanding Syntax
4. Javascript understands some
5. Understanding syntax in C + +

In order to achieve this, of course, the simplest and fastest is the best. Considering that the university has been using C #, do you want to use it? But found that the OSX platform can only use mono, but also to be re-familiar. Java implementation is also unpleasant, from the time needed to consider. JavaScript is not familiar, it seems that you can use node. js to write (Atom is the use of it). Not familiar. C + + has not been used for many years, and, to implement a lot of code, especially troublesome. Consider the previous period of time just before codecademy learn grammar, take it to practiced hand it.
ok, identified with Python. Follow-up is how to request the network, parsing page HTML tags, extract download links, download files. Although do not understand how these in Python is implemented, but the process is determined, according to the process to the site to find out, here do not study the principle, realize the function.
Next is a variety of search engines, Google, Baidu can also (different engine focus is not the same). Don't forget what the purpose is, search for the relevant information.
OK, after the search, determine the request network download Web page with requests, parsing HTML with beautifulsoup, extract download link BeautifulSoup, Download the document (the code for a download file found in StackOverflow).
and then you put them together. The code after the combination is as follows:

#file-name:pdf_download.py __author__ = ' rxread ' Import requests from BS4 import BeautifulSoup def download_file (URL, Index): Local_filename = index+ "-" +url.split ('/') [-1] # NOTE the stream=true parameter r = requests.get (Ur              L, Stream=true) with open (Local_filename, ' WB ') as F:for Chunk in R.iter_content (chunk_size=1024): If Chunk: # filter out keep-alive new chunks F.write (chunk) F.flush () return L Ocal_filename #http://ww0.java4.datastructures.net/handouts/root_link= "http://ww0.java4.datastructures.net/       handouts/"R=requests.get (Root_link) if R.status_code==200:soup=beautifulsoup (r.text) # print soup.prettify () Index=1 for link in soup.find_all (' a '): New_link=root_link+link.get (' href ') if new_link.endsw ITH (". pdf"): File_path=download_file (NEW_LINK,STR (index)) print "Downloading:" +new_link+ "   "+file_path index+=1   print "All download finished" Else:print "errors occur." 

You can download all PDF documents locally by running the following code.

Python pdf_download.py

3. Optimization

More than 30 lines of code, all done, it is simple and clear, sure enough to do Python for some script task is still good. It downloads 41 of documents.
The first download from the document did not have a serial number, so I do not know the time to see, so I gave the file name preceded by a serial number.
Other optimization sections can be referenced as follows:

1. Consider that some of the exceptions to the function are not handled and need to be processed later.
2. Functions are not fully encapsulated, the downloaded file type support is not much, this can be extended according to their own needs.
3. It may be possible to download fewer files, but with more files, it is necessary to use multiple threads (the right amount) or a thread pool to download, which speeds up the download.
4. Some of the wording may not conform to the Python syntax specification, of course, written and not written is already 0 and 1 difference.
5. Other details, such as PDF may be in uppercase PDF.

4. Appendix

1. "Data structures and Algorithms in Java" (Michael T. Goodrich, Roberto tamassia) Download http://bookzz.org/or http://it-ebooks.i nfo/
   The following two sites are good book download site, conditional or buy a copy of the book to support the author it.
   Generally I will first download the ebook to see, appropriate to buy a paper version.
2. Getting started with Python syntax Http://www.codecademy.com/zh/tracks/python

Above, that is the case.

This article from Rxread's Blog, welcome reprint, reprint please specify.
Welcome to discuss the exchange together.

Python crawls all PDF documents in a single page