Necessity: numpy,argparse,wget,arxiv,pdfx 
python 2.7

download them by :

pip install argparse wget arxiv pdfx

>>> arxiv_search_download.py
Search or download the arXiv or even aps papers!
													
Usage-1, find someone's paper (cat can be str-el, mtrl-sci, supr-con or all), eg:
python arxiv_search_download.py --aut="Piers Coleman" --num=50 --cat=all
													
Usage-2A, find some topic, eg:
python arxiv_search_download.py --key="machine learning" --num=50 --cat=all
Usage-2B, find some topic, and shown abstrct, eg:
python arxiv_search_download.py --key="machine learning" --num=50 --cat=all --abs=1 
													
Usage-3, find someone's paper and download to your directory, eg:
python arxiv_search_download.py --dir="ImpuritySolver" --key="Impurity Solver" --cat=all  --num=50 --dwn=1
													
Usage-4, get the latest papers (and even download it!), just run
python arxiv_search_download.py --num=10
python arxiv_search_download.py --num=10 --dwn=1 --dir="new" 

Usage-5 get latest abstracts and search your keywords
python arxiv_search_download.py --abs=1 --cat=all --num=1000 > abstracts_latest_1000
grep "supercondu" abstracts_latest_1000 | head -n 10
													
For more details, please refer to 
https://arxiv.org/help/api/user-manual or https://github.com/lukasschwab/arxiv.py
Developer : [email protected], based on arxiv interfaces.


>>> get_links_download_in_pdf.py

Search or download the refs as links in a pdf file (not works for scanned pdf or papers without link)

>>>  download_aps_citedby.py
(please see this script and find the necessary modules in "import ..." and fix it).
Search or download the citing papers for an aps cite-by link, e.g.
python download_aps_citedby.py --url="https://journals.aps.org/prl/cited-by/10.1103/PhysRevLett.123.090605" --dir="prl_123090605"




Note: Please let the scripts sleep longer to avoid punishment (time.sleep function inside) !!!