primo | primo轮子中心，为开发效率翻滚。（Java、Spring）一些自动生成代码的插件，缓存、限流、接口等快速接入的中间件也会逐渐弄起来 | Object-Relational Mapping library

This modified awk script from my last answer should work for you:

You have an error in your FractionCollection.put method. When you add a fraction to the innerContainer array you increment logicSize, but it's just a local variable. The actual this.size remains unchanged. Instead of

205|206 result will be 1

I would suggest passing tesseract every row of text as separate image.
For some reason it seams to solve the decimal point issue...

• Convert image from grayscale to black and white using cv2.threshold.
• Use cv2.dilate morphological operation with very long horizontal kernel (merge blocks across horizontal direction).
• Use find contours - each merged row is going to be in a separate contour.
• Find bounding boxes of the contours.
• Sort the bounding boxes according to the y coordinate.
• Iterate bounding boxes, and pass slices to pytesseract.

Here is the code:

You are correct in trying to use 3d party Http client rather than doing it with HttpURLConnection class. What you can do is try some other HTTP clients available. Popular ones would be Apache Http Client or OK Http client. However, I may suggest my own Open Source MgntUtils library with Http client as well. It may be not as extensive as other library but is very simple and geared for sending requests multiple times to the same URL. In case you want to use my library your code would look something like this

You works with every line separtelly so you get only X. To get Y you need to enumerate lines:

You can use selectors to access elements. Here you want to select by tag name. Two ways to get the element you want:

Your list code seems OK.

Problem is with how you use it - you never update the variable pl. So do something like that:

I'd suggest looking at the network tab in chrome developer tools and then submitting the request to load up the URL ... you'll see a lot of requests going back/forth.

Two that seem to contain relevant content are:

https://merhav.nli.org.il/primo_library/libweb/webservices/rest/primo-explore/v1/pnxs?blendFacetsSeparately=false&getMore=0&inst=NNL&lang=iw_IL&limit=10&newspapersActive=false&newspapersSearch=false&offset=0&pcAvailability=true&q=any,contains,%D7%94%D7%90%D7%A8%D7%99+%D7%A4%D7%95%D7%98%D7%A8&qExclude=&qInclude=&refEntryActive=false&rtaLinks=true&scope=Local&skipDelivery=Y&sort=rank&tab=default_tab&vid=NLI

which requires a token to access token which comes from:

https://merhav.nli.org.il/primo_library/libweb/webservices/rest/v1/guestJwt/NNL?isGuest=true&lang=iw_IL&targetUrl=https%253A%252F%252Fmerhav.nli.org.il%252Fprimo-explore%252Fsearch%253Ftab%253Ddefault_tab%2526search_scope%253DLocal%2526vid%253DNLI%2526lang%253Diw_IL%2526query%253Dany%252Ccontains%252C%2525D7%252594%2525D7%252590%2525D7%2525A8%2525D7%252599%252520%2525D7%2525A4%2525D7%252595%2525D7%252598%2525D7%2525A8&viewId=NLI

.. which likely requires the JSessoinId which comes from:

https://merhav.nli.org.il/primo_library/libweb/webservices/rest/v1/configuration/NLI

.. so in order to replicate the chain of calls you could use JSoup to make these (and any other relevant) HTTP GET requests, pull out the relevant HTTP headers (typically: session, referer, accept and some other cookie values potentially)

Its not going to be straight forward, but you're essentially looking for a url on the page in one of the JSON responses from one of the network requests:

Once you know which request you want to recreate, you just have to work back up the list of requests and try to recreate them.

This one is not an easy one and would require a lot of time to recreate - my advice if you're going to attempt it, forget trying to parse HTML, try to rebuild/recreate the chain of 3 or so HTTP requests to the back end to get the relevant JSON and parse that. You can often pick apart the website but this ones a big job