H2O flow - save to html/pdf as a report - jupyter

Is there a way to save h2o flow cells after training, in form of a HTML or PDF report? I would like to achieve similar result to Jupyter Notebook, full of figures, tables etc. - a static overview of what happened. "Save flow" option saves static text, but removes figures and pictures.
So far I've tried to use primitive approach - just "save as HTML full page" in Chrome/Firefox/Safari/Edge, but the result is either empty page, or just a first cell with missing graphic elements.
I was using static Jupyter notebooks as a form of report for my Clients, and I'd like to use H2O Flow in a similar way.
Thank you in advance

The best solution is to do a "print to pdf" from the browser. On a Mac, you'd go to File -> Print -> click on the PDF button -> Save to PDF.
Here's an example of the PDF output for a simple AutoML run on the iris dataset.

Related

BIRT Report Designer: Table of Contents (TOC) with page number and hyperlink

I'm using BIRT report designer v4.8 and trying to build a table of contents(TOC).
I know that BIRT has a builtin Table of Contents in Properties, but that kind of TOC can only be viewed in the Web Viewer by clicking the "Toggle table of contents" and be showed up on the left side of the report.
What I want to do is to create a TOC that is part of the report, which means the TOC is directly showed up as a section in the report without clicking the "Toggle table of contents", and also can be exported along with the report.
For example, here is a sample of TOC, suppose we have four sections in the report.
The challenge I'm facing now is calculating the page numbers for each section and then add a hyperlink for each section so that the page will jump to that section when a user clicks it from the TOC. Does anyone know how to get this functionality? What kind of scripts do I need?
This is not possible with BIRT alone.
A possible solution is depending on the output format.
With PDF and a good knowledge of iText, it should be possible to create a TOC in a post-processing step.
I'll sketch the idea here (assuming that the TOC will fit on one page):
With BIRT, generate your invisible TOC as before. On the first page, leave enough room for the TOC.
The rest of the work will happen after BIRT has generated the PDF. BIRT contains iText and iText can examine and modify existing PDFs.
In PDF speech, the TOC is called "outline". With iText, examine the outline. This will give you the data you need for your visible TOC and can be transformed to an array of triples of the form (toc_level, toc_text, page_no).
Then you can create a new PDF with the help of iTexts PDFStamper class: Take the existing PDFs and render your list of triples on page 1.
This will cost you at least 1-2 days of Java development work, however.

Reading Data from excel sheets in Katalon studio

How to read data from excel sheets(sheet 1,sheet 2...etc) in Katalon Studio.
I am able reading the data from one sheet.But I am facing the difficulty with multiple sheets.
Without knowing exactly what the problem you face is, here are my first suggestions to check:
You do have to create different Test Data files for each sheet in your file that you want to use. Two Test Data files for the same spreadsheet file
In your new Test Data inside Katalon, just select your file and then you can select the different sheet. Sheet selection options
If you can't see a sheet in the drop down on your Test Data page (perhaps you just added it to the spreadsheet) refresh your project, or at the very least refresh the Test Data. (ctrl-F5 to refresh the entire project, highlight the Test Object and then right-click and select or hit F5 to refresh the Test Data)
Hope this helps.
You cannot switch excel sheets from Katalon, at least natively. Maybe with some custom programming.

Automation of clicks on webpage using Matlab

So I'm looking for a way to systematically access data from a website. This data is updated every 15 minutes or so, and is generated through a datamart system that makes custom reports following several input parameters: the desired date interval, the specific dataset.
All these parameters require me to click on some specific buttons; I was wondering if it would be possible to automate these click inputs using Matlab (or something else if need be), to retrieve the data and treat it automatically.
Thanks in advance!
I suggest you take a look at http://www.autohotkey.com/. This is a great tool which allows for the automatic clicking on any window (including a browser page) under Windows. It even will let you "search" your screen for pixel images and then click on those images. This would allow you to make a very small bmp file of the link you would like to click on, and then your script can search your page and click directly on the link.
As far as getting data into matlab I'm not exactly sure of the best way to do this, but you might consider saving the html of the page, and then parsing that from matlab.

Printing the contents of a Form

I have a form with some comboboxes, a datagrid and an image. My ultimate goal here is: When the user clicks the save button on a menustrip, it will send this data to a printer (preferably to a PDF one) and print the file ready to be mailed to our clients. I tried the code provided by:
http://msdn.microsoft.com/en-us/library/aa287529%28VS.71%29.aspx
But it sent me to OneNote which then showed me an empty file. I also tried Printform and PrintDialog which didn't bring me much luck either. Any suggestions?
edit: I configured it to print to pdf although it now prints only half the form (cuts it in half of width), and if the datagrid is Scrollable it only prints the data that is showing on the screen. Like this:
You cannot print scrolled data using graphics object. even you cannot create your own design using graphics object. Also, you will get some problem in creating PDF and managing multiple pages. So, don't make it more complicated just use reporting library. You have two choices to print form data into pdf file.
Microsoft Reporinting
Crystal Report
But, I would like to prefer to use Crystal Report to Print any document. It is more flexible than Microsoft Reporting. You can export that document into PDF, WORD, EXCEL or Direct to Printer.

How to export data from Chrome developer tool?

Network analysis by Chrome when page loads
I would like to export this data to Microsoft Excel so that I will have a list of similar data when loaded at different times. Loading a page one time doesn't really tell me much especially if I want to compare pages.
if you right click on any of the rows you can export the item or the entire data set as HAR which appears to be a JSON format.
It shouldn't be terribly difficult to script up something to transform that to a csv if you really need it in excel, but if you're already scripting you might as well just use the script to ask your questions of the data.
If anyone knows how to drive the "load page, export data" part of the process from the command line I'd be quite interested in hearing how
from Chrome 76, you have Import/Export buttons.
I was trying to copy the size data measured from Chrome Network and stumbled on this post. I just found an easier way to "export" the data out to excel which is to copy the table and paste to excel.
The trick is click Control + A (select all) and once the entire table will be highlighted, paste it to Microsoft Excel. The only issue is if there are too many fields, not all rows are copied and you might have to copy and paste several times.
UPDATED: I found that copying the data only works when I turn off the filter options (the funnel-looking button above the table). – bendur
Right-click and export as HAR, then view it using Jan Odvarko's HAR Viewer
This helps in visualising the already captured HAR logs.
I came across the same problem, and found that easier way is to undock the developer tool's video to a separate window! (Using the right hand top corner toolbar button of developer tools window)
and in the new window , simply say select all and copy and paste to excel!!
In Chrome, in the Developer Tools, under Network, in the Name column, right-click and select "Save as HAR with content". Then open a new tab, go to https://toolbox.googleapps.com/apps/har_analyzer/ and open the saved HAR file.
Note that ≪Copy all as HAR≫ does not contain response body.
You can get response body via ≪Save as HAR with Content≫, but it breaks if you have any more than a trivial amount of logs (I tried once with only 8k requests and it doesn't work.) To solve this, you can script an output yourself using _request.contentData().
When there's too many logs, even _request.contentData() and ≪Copy response≫ would fail, hopefully they would fix this problem. Until then, inspecting any more than a trivial amount of network logs cannot be properly done with Chrome Network Inspector and its best to use another tool.
You can use fiddler web debugger to import the HAR and then it is very easy from their on... Ctrl+A (select all) then Ctrl+c (copy summary) then paste in excel and have fun
I don't see an export or save as option.
I filtered out all the unwanted requests using -.css -.js -.woff then right clicked on one of the requests then Copy > Copy all as HAR
Then pasted the content into a text editor and saved it.
I had same issue for which I came here. With some trials, I figured out for copying multiple pages of chrome data as in the question I zoomed out till I got all the data in one page, that is, without scroll, with very small font size. Now copy and paste that in excel which copies all the records and in normal font.
This is good for few pages of data I think.
In more modern versions of Chrome you can just drag a .har file into the network tab of Chrome Dev Tools to load it.
To get this in excel or csv format- right click the folder and select "copy response"- paste to excel and use text to columns.
You can try use Haiphen, which is a chrome extension that allows you to analyze network traffic and what API calls a web application is making.