HNHacker News
TopNewBestAskShowJobs

saradhi

269 karma · joined January 15, 2020

I had problems extracting tabular data from images and PDFs, so I created a Document Structure Detection API - https://extracttable.com

saradhi[@]extracttable

submissionscomments
saradhi··on Show HN: Draw and extract tables from PDF reliably
Reliably* - coming with limitations

  1. Text pdfs only 

  2. Template is needed
From what I've spoken with 50-60 paid customers for a service I maintain, the most common concerns I listen are

  1. No, We get scan pdfs as well 

  2. No, We deal with lot of customers, cannot template all
saradhi··on Show HN: Pdf.js Express – PDF annotation, e-signatures, and form filling
"can" - easier said than done, especially tables.

I don't see it in the Demo. Did I miss it ?!

saradhi··on What's so hard about PDF text extraction?
I understand how the whole thing interpreted as, and I agree with you. I'll do a note as suggested. Thanks for guiding and the link too.

I would like to get the author's comment on "tried to contact you and was ignored", as I was the one who emailed yesterday.

saradhi··on What's so hard about PDF text extraction?
- No, No, Zero.

"commercialize the original author's work with the author" - No, but let me highlight this, any extraction with tabula-py is not commercialized - you can look into the wrapper too :) or even compare the results with tabula-py vs tabulaPro.

Copying the TabulaPro description here, "TabulaPro is a layer on tabula-py library to extract tables from Scan PDFs and Images." - we respect every effort of the contributors & author, never intended to plagiarize.

I understand the misinterpretation here is that we are charging for the open-sourced library because of the name. We already informed author in the email about unpublishing the library, this morning, I just deleted the project and came here to mention it is deleted :)

saradhi··on What's so hard about PDF text extraction?
"pencil written notes that bleed outside cell borders, curve around borders, etc.? Stamps and symbols (watermarks) across tables?"

"The Worst Image" is a close match to that, except it is a print.

Regarding increasing degradation - as stated above, the OCR engine is not proprietary - we confined ourselves to detect the structure at this moment, and started with the most common problems.

saradhi··on What's so hard about PDF text extraction?
William, the intention of "TabulaPro" is to give the developers a chance to use a single library instead of switching ExtractTable for images and tabula-py for text PDFs.

What do you recommend us to do, to not make you feel we made a dick move.

TIA

saradhi··on What's so hard about PDF text extraction?
Interestingly, I was doing the similar stuff for 3 years to a US company. Curious, is your client a legal tech company? Mine was.

The experience helped me to roll out an API, as https://extracttable.com, for developers.

OCR tricks? Assuming post processing dev stuff - may I know your OCR engine. We are supported with Kofax and openText along with cloud engines like GVision as a backup.

saradhi··on What's so hard about PDF text extraction?
I'm an ML engineer, worked as a part time data engineer consultant for a medical lines/claims extraction company, for 3 years, which majorly involved in extracting the tabular data from the PDFs and Images. Developer rules or parsers as such is JUST no help. You end up creating a new rule every time you miss the data extraction.

With that in consideration, and the existing resources are little help especially on skewed, blurry, handwritten and 2 different table structure in the input, I ended up creating an API service to extract tabular data from Images and PDFs - hosted as https://extracttable.com . We cared it to be robust, average extraction time on images is under 5 seconds. On top of maintaining accuracy, A bad extraction is eligible for credit usage refund, which literally not any service offer it.

i Invite HN users to give it a try and feel free to email saradhi@extracttable.com for extra API credits for the trail.

saradhi··on Ask HN: How do you decide to accept new job offer or not
In the order:

  sponsorship - because I'm an immigrant

  Tech stack - I spend most of my after office hours in doing new projects, effectively, I like to use the skills learned in the job. So an advanced tech and freedom to experiment with tech is always a top preference

  sustainability - a continuous revenue-generating company or dependent on new sales every time

  motivation - be it a side project or office work, I enjoy working on the problems that are pointed out - greater the impact, more the involvement

  hierarchy - from my experience, too many levels above you is a problem not just in career growth but also product implementations, which kills the pace and transparency

  compensation - a 20% hike is the least I expect
saradhi··on China to inject $174B of liquidity on Monday as markets reopen
Ask HN: Should I be worried about the PANIC around the current situations, will they act as a rescission trigger? Can I assume a proportional impact on crypto-currency market as well?

More importantly, I'm considering a move from a Bn$ organization to a safety concerned mobile-app product startup, How to access whether a start-up can survive if the situation end-up on the negative side, in the next 6 months.

saradhi··on Show HN: Clips.live – Live video clips of every NBA play
Unsure why ICC was involved in your case. Mine was pretty much with BCCI and Sony, (the channel which owned broadcasting rights of IPL at that time). They pretty much had a very neat clause, which I should have taken a note upfront, nevertheless, the experience gained confidence in me to build applications. :)
saradhi··on Show HN: Clips.live – Live video clips of every NBA play
Did you invest in time to check if you are free to publish clips, especially with "clips.live" as a watermark. I once did the same in another sport, cricket, back in 2013, ended up killing it because of a cricket board and its broadcasting violations note over an email.
saradhi··on Software engineers complaining about a $450k salary?
if only social service and society impact was measured for the pay, me, my other friends and may be you would have been homeless
saradhi··on Ask HN: Bootstrapped founders, how do you conduct market research?

  check, if the product/service solves my purpose

  quick search for users# like me (profession or personality)
saradhi··on Table Detection and Extraction Using Deep Learning
Did you try https://extracttable.com

The mentioned service is not perfect either. There are always limitations, minimizing is the key.

P.s: I work with the team at extracttable

saradhi··on Ask HN: Early in career – is it a bad idea to change jobs after six months?
> Now I'm drowning in work, with no one to give me priority or direction.

When you feel that "drowning", probably you are not enjoying the work load, which is quite important for the happy-you. I understand you are in the early days, did you try to make your decisions on the priority? Sometimes, you may be better, who knows. Give it a try.

> The work I am asked to do is important and I don't feel comfortable telling individuals high up in the subsidiary no.

This is the situation you've to face in your life, every day. Saying "no" is perfectly "yes", be bold enough to say "no". But just give a thought on the after-effects, if those do not put you in trouble, it is perfectly alright.

Good luck.

saradhi··on Ask HN: How to stop "demo" abuses
Changed the title. Also, added the info.

I'm looking for help in limiting the users after certain number of trails per day. Currently, I'm relying on a counter that is referred from the "localStorage" of the page

saradhi··on Ask HN: Incorporate with Clerky vs. Atlas vs. Gust in 2020?
That’s definitely competitive in pricing. Thanks for the mention.

I'm curious to know more about "-Special Offers from our partners" that comes in the package. Do you have any idea on it. Couldn't find details on it.

Edit: contacted support on chat, offers are for tax filings and maintenance fees. No cloud /partnered credits

← PreviousPage 3 of 3