ckreibich / scholar.py

A parser for Google Scholar, written in Python
2.11k stars 777 forks source link

Any front-end UI of scholar.py? #99

Open reflect9 opened 7 years ago

reflect9 commented 7 years ago

Hello folks,

I have always been wondering if there exists a powerful environment for searching articles on Google Scholar. I think this repository provides all low-level features, but it does not have high-level application features such as custom aggregated visualization, etc.

I would like to contribute if there's any open-source project for building front-end of this project. If not I would like to start my own. Any advice?

best regards, Tak

s-m-e commented 7 years ago

I am not aware of any GUI front end for scholar.py (but I might simply be wrong, so do NOT rely on my answer). Other software uses Google Scholar in a similar manner and does have a front end, e.g. Zotero. Zotero extracts what it calls "representative phrases" from pdf files and uses Google Scholar for finding citation information on those files by searching for the "representative phrases". It's a nice way of document recognition. The program is written in JavaScript and handles captchas very well by presenting them to the user. Its Google Scholar code is not particularly easy to use (i.e. call from another program) or port to another language. Therefore I have been considering to develop something similar to what you suggested based on scholar.py plus document recognition - in Python, adding this functionality to the Python ecosystem. Keep me in the loop if you're actually interested in going down this path. I'd be interested to help.

eric-burel commented 6 years ago

Hi, I am a frontend dev and I write a lot of technical studies for companies, so I am also a recurrent user of Scholar. Existing softwares such as Zotero seem powerful, but in terms of UX... I mean, why a powerful software should always be ugly as hell ? We are in 2017 now, writing good looking ergonomic UI that works on any platform is actually easy.

So I'd be deeply interested in such a project, with the goal of helping user to write state-of-the-art and generate bibliography quickly. Vue.js would be a great candidate because it's awfully easy to develop beautiful webapps with it. We would simply need a minimal python server to run scholar.py. As I am learning Elm, I may also consider this solution, because it is very well structured and scales well, but this is more for my own educationnal purpose ^^ And then of course we could simply build a Python app.

Also, one needs to consider that Google Scholar technically does not authorize such a scraper, even less with its own frontend, so the UI should be as respectful to Scholar as this package is.

wasified commented 6 years ago

I have felt the same need, and I already have a project surrounding this theme of a "modern" reference manager. I've used Zotero and the lot too, but I've felt that the UI isn't that great and there's this constant need to manage multiple screens (which is counter-productive). The ideal solution is to embed these tools in the writing environment.

I started building it around scholar.py, but scholar.py is too slow so I wrote my own module to make the process faster. It mimics the actual website, so it's as respectful as Scholar itself. It's built around Python, and uses HTML + CSS for places where richer content is needed. If anyone is interested I can delve on the tech stack a bit more.

Right now, I'm adding in keyboard shortcuts (auto-fill for citations) like typing in @ and then typing the authors name will automatically add in the proper citation in your document. It feels really natural to just add in a citation with a single keystroke. Personally, I'm just tired of jumping through different interfaces to get stuff done.

I'm developing it alone right now, and the thought of working with other people on it really excites me. You can find the current release here: www.slate.ink. Do check out the video walk-through. Criticism is more than welcome.

On Tue, Nov 7, 2017 at 2:40 PM, Eric Burel notifications@github.com wrote:

Hi, I am a frontend dev and I write a lot of technical studies for companies, so I am also a recurrent user of Scholar. Existing softwares such as Zotero seem powerful, but in terms of UX... I mean, why a powerful software should always be ugly as hell ? We are in 2017 now, writing good looking ergonomic UI that works on any platform is actually easy.

So I'd be deeply interested in such a project, with the goal of helping user to write state-of-the-art and generate bibliography quickly. Vue.js would be a great candidate because it's awfully easy to develop beautiful webapps with it. We would simply need a minimal python server to run scholar.py. As I am learning Elm, I may also consider this solution, because it is very well structured and scales well, but this is more for my own educationnal purpose ^^ And then of course we could simply build a Python app.

Also, one needs to consider that Google Scholar technically does not authorize such a scraper, even less with its own frontend, so the UI should as respectful to Scholar as this package is.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ckreibich/scholar.py/issues/99#issuecomment-342427566, or mute the thread https://github.com/notifications/unsubscribe-auth/AGEL-S6KOwdvctTVCJcXnmo3qs4ArXHuks5s0CV0gaJpZM4PeMXV .

russelljjarvis commented 4 years ago

Hi, I made a lot of progress with using streamlit and scholar.py together. https://github.com/russelljjarvis/ScienceAccess It's a work in progress. Actually gets stats on texts scraped from GScholar.r

Streamlit almost writes itself. No messing around with html elements necessary.