Multiple pdf download qith mutiple steps

[bobybonillamx]

Hi,

I have an issue with a flow in desktop

I need to download like 19,000 pdf's from a web page, but the web page has 3 steps to get the pdf download button

i've tryed with web page data extrator and recording the steps but the fisrt button in the page get a blank text input at beginning but when i press it get a text with the specific name of the file (not a overall button text)

Si i tryed  to make an apply to each in the flow but it doesn't detect the rest of the buttons because they have an specific title linked to the pdf name 

This is how the web page presents the buttons, the 3 dot button it's the firs one to click

and so on it doesn't have pagination it's just a scroll list limited by 500 files each time i ask for them

that first button the UI detector of PA doesn't get the title, so it don't make the table un the web extractor

When you click in that 3 dot button displays the options and ther i need to click in the print one

When i click in print manifest i get another option in a pop up mode

and there i need to click in the coversheet button to get the pdf file

This i need to repeat the action with each of the files but i can't make the flow work by it self to move on with the next in the list

any advise?

Thanks

[Pavel_NaNoi]

This is just an assumption but the website three dots are probably identified by an id such as div:eq(n) or button:eq(n) where n is any value from 0 - the number of three dot drop down buttons.

So, in this case what'll work is:

• Extract data from webpage all three dot dropdown buttons
• Start loop (not a for each, just a normal loop) that goes from 0 to the number of rows extracted from page
• Click link on webpage, with a custom selector, go into text editor and put in the selector from Extract data from webpage action, but with a variable inside whatever CSS element is responsible for selecting the three dot dropdown, so the aforementioned div:eq(n) or button:eq(n) above, might even be something else completely, just play around with it.
• Click link on webpage but now press the print button
• Wait for website content / Wait for window content (depends on the popup, but just wait for that element)
• Click link on webpage / Click UI Element (again depends on the popup) onto the print manifest coversheet button
• End of loop

That should do it I think.

[momlo]

Hi @bobybonillamx 

I will not answer directly to your questions but ask if the below approach would be possible in your case.

1. Can you obtain the list of those 15k documents in a form of a list/excel etc?

2. Does the "print manifest" button has a URL that you can capture? Is this URL unique per file?

3. If the URL from point 2 is unique per document, do you have all the details used to build the URL included on your list form point 1?

And now, if answers to all points are "yes", what you could try to do is:

1. Get the list of documents

2. For each document build URL and open it, then click the button, do whatever you need to do afterwards, end loop iteration

[bobybonillamx]

Hi, sorry for the delay

the 1st point "

• Extract data from webpage all three dot dropdown buttons"

i assume it's like a title because i don't understand if it's a instruccion how to do it,

i have been looking in the CSS to play arround with the id of the button but the 3 ID that i see o nothing, i asume it's because the espacial text that embed in the button when i click on it

i'm trying to replicate an understand all that you write me but i think i'm missing details from it 

i will tell you if i achive something new thanks!

[bobybonillamx]

Hi!

1. NO it just shows as is in the image is a scroll down list 

2. YES i'ts unique but it only "exist" when i click in the 3 dots before that is not present 

[momlo]

ok, i understand the link is not shown until you click the dots, but can you test if you do something like:

1. Open the list of documents

2. Click the dotts, right click the button with link and copy the link

3. Go back to the list of documents, so it is like you never did (2). Then paste the copied URL into browser and see where it takes you

[Pavel_NaNoi]

Sorry, to be more specific,

use the "Extract data from webpage" action, then go to your webpage a little console should popup when you do, right click on the three dots dropdown on the webpage to extract its contents, and extract text for example, then go onto a different dropdown, do the same, this should then extract all dropdowns on the webpage and you'll see outlines around them, once that is done, go into the "advanced options" in the console, and copy the CSS selector you see, copy both the base and the target selectors, and put them together as such:

then use that as the custom selector I mention in the next step,

Let me know if anything else doesn't make sense.

[bobybonillamx]

Hi,

i now understand but it doesn not make that, i have used that console before but for some reason that 3 dots doesn't seem to work with that, look

When the estract dat afrom web console is open it looks like this:

it doesn't "read" the name so the list it's in blank and is where i can't move on

but i now see what you are talking about the selectors i got this in the advanced options

So i will use like this if i'm undertanding well:

html > body > cbptm-root > cbp-root > div > cbptm-manifest-list > div > div:eq(1) > cdk-virtual-scroll-viewport > div:eq(0) > cbptm-view-cards > mat-card > mat-card-header > div > mat-card-title > div > div:eq(6) > button:eq(1)

Let me try with that 

[bobybonillamx]

Ok so i've try this but i keep getting errors

it seems like the popup changes every time it opens for some reason the flow works now with the 3 point and the print manifest button but the coversheet i can't make it click

i try the web extract thing you said in the advanced options but it shows as single value

[Pavel_NaNoi]

Its supposed to be a single value for that button in particular, if it "changes" each time you need to make it more generic, try making the single value selector from html > body > .... to just button:contains(Print Manifest Coversheet).

Also your code will open up 300 microsoft edge instances,

please put that outside the loop, and also make the 300 loop index the datafromwebpage.rowscount instead.

[bobybonillamx]

ok so, fter this week

i finally don't get errors in the button pushes

but the loop is now working and only download the 1st document over and over again

i noticed that the 2nd button the "print manifest" in the drop down menu has this

and that cdk-overlay changes to 7 or 20 or 61 without any rational list or arrange in the next button is like this

div[Id="cdk-overlay-61"] > div > div > button:eq(1)

The 3 dots button has this example

in th first button is:

> body > cbptm-root > cbp-root > div > cbptm-manifest-list > div > div > cdk-virtual-scroll-viewport > div > cbptm-view-cards > mat-card > mat-card-header > div > mat-card-title > div > div > button > span > i:eq(1)

but in the second button is:

> body > cbptm-root > cbp-root > div > cbptm-manifest-list > div > div > cdk-virtual-scroll-viewport > div > cbptm-view-cards > mat-card > mat-card-header > div > mat-card-title > div > div > button > span > i:eq(3)

in the third button is:

> body > cbptm-root > cbp-root > div > cbptm-manifest-list > div > div > cdk-virtual-scroll-viewport > div > cbptm-view-cards > mat-card > mat-card-header > div > mat-card-title > div > div > button > span > i:eq(5)

i suppose the 3 dot buton has a +2 modification but i don't kno how to mark that in the loop

The last button that is the coversheet seems to work well with the Span class but i can't verify further 

[Pavel_NaNoi]

To mark that in the loop, just do this:

- Create a new variable outside the loop, name it ButtonIndex or something and set it to 1

- In your selector now change it to the following:

> body > cbptm-root > cbp-root > div > cbptm-manifest-list > div > div > cdk-virtual-scroll-viewport > div > cbptm-view-cards > mat-card > mat-card-header > div > mat-card-title > div > div > button > span > i:eq(%ButtonIndex%)

- Finally, put an "Increase variable" action at the end of the loop and tell it to increase the ButtonIndex variable by 2.

That should work.

[bobybonillamx]

Wow i-m getting close to it now

I have got the flow to push all the 3 dot button that are visible in the screen and also download the PDF succesfully

But now when it whet to the bottom of the list (without scrolling down) it stops to work and send an error that doesn't find the next button

i suppose it's because the button is not visitble in the page

i've tryed to scroll down with java and focus on text with out success

the div look like this:

> body > cbptm-root > cbp-root > div > cbptm-manifest-list > div > div > cdk-virtual-scroll-viewport > div

[bobybonillamx]

ok so new issue jaja

the 3 dot button that increase value +2 in the list is defined by the position in the window

so if the 9th button in the list is at the top of the list (because the scroll) it has de ID 3 again en the 10th 5 and go on

so when i scroll down the id moves with the scroll

i'm getting frustrated

[Pavel_NaNoi]

That's one annoying webpage, christ who writes evil HTML code like that? alright try the following:

In your loop do an if statement that checks when it hit the last ID on the screen via a number e.g in the case above that you've provided: If ButtonIndex >= 20 (because the ID of the 8'th button would be 20) then: Set variable ButtonIndex = 1 and then scroll down.

This will reset it (I'm assuming you've found a way to scroll down reliably) and start the next loop looking at the first ID.