When I run the code in the example, I get the following error:

AttributeError: wdStory

May I know how to fix this?
Also, it would be nice if error messages would be more informative.

Thanks in advance!

• Our CI should test against pip install rpaframework probably nightly and against some of the biggest versions to catch problems where floating subdependencies break user setup
• Do scripted analysis to figure out which of our dependencies lack pinnings for their dependencies major dependencies and figure out if fixing this case by case is a good solution

Hello,

this might be related to #23

Running rpaframework 5.3.3

I want to move a file on a network share:

Example not working
    RPA.FileSystem.MoveFile    ${source}    ${destionation}    overwrite=${True}

Example working
    OperatingSystem.Move File    ${source}    ${destination}

It works with OperatingSystem library, but not with RPA.FileSystem due to Permission constraints.

Regards,
Markus

Hi,

the imap port specified in ImapSmtp Library is never used. Imap connections always go against port 993.

I think, this line where the connection is created, is missing the port:

Kind regards,
Markus

In the below example Excel and Calculator are shown in the open applications list, but not the Notepad which opens .txt file.

${app1}    Open Application   Excel
${app2}    Open Executable    calc.exe  Calculator
${app3}    Open File          /path/to/myfile.txt
&{apps}    Get Open Applications

Initial version should be able to do web scraping for elements on the page, especially tables using BeautifulSoup package.

Could be added as a new separate library or feature included into current libraries like RPA.Browser and RPA.Desktop

For future versions add support for scraping tables from images

Sometimes when an RF script fails, or teardowns are not executed for some other reason, a chromedriver process stays running in the background. This creates various issues because it prevents cleaning up folders and pollutes the user's machine with stray processes.

It would be possible to have some hooks on exit and make sure the created webdriver process is killed, but sometimes it's wanted that the browser should stay running after a script exists. Need to figure out some way to fulfill both requirements.

I dropped OperatingSystem Library and used RPA.FileSystem instead. Now I run in to an odd permission error when calling Normalize Path:

I have no idea, why permission rights are required or what OperatingSystem does differently.

Currently PIL / pyscreenshot only captures the active monitor, which limits the usability of template matching in multi-monitor environments. RPA.Images should be improved to capture images from all available display devices, and the resulting coordinates should map correctly with RPA.Desktop.Windows.

Hello,

following example: I have a dictionary mapping worksheet names and tables. Those tables include headers. When running this keyword, headers are missing in the final worksheet.

Save values in '${ausgabedatei}'
    create workbook    ${OUTPUT_DIR}/${ausgabedatei}
    FOR    ${name}    ${tabelle}     IN    &{PRODUKTWERTE}
        create worksheet    ${name}    ${tabelle}
    END
    save workbook
    [teardown]    close workbook

Regards,
Markus

As the title suggests, if feasible to implement the Selenium library should have keywords for inspecting network traffic, e.g. reading response headers. This would bring some parity between it and Playwright.

See robocorp/robotframework-lsp#199

to be included in the RPA.Cloud.Google library

this also "forces" to implement support of OAuth2 authentication using credentials.json

keywords

• Init Apps Script Client
• Run Script

https://developers.google.com/apps-script/api

Chromedriver should be in the PATH but Open Available Browser keywords opens Safari instead of preferred Chrome.

https://forum.robocorp.com/t/browsernotfounderror/96

20200813 20:31:01.984 [WARN]: Importing test library "RPA.FileSystem" failed

Traceback (most recent call last):
list index out of range
File "/usr/local/Cellar/python/3.7.7/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/robotide/spec/librarymanager.py", line 82, in _fetch_keywords
return get_import_result(path, library_args)
File "/usr/local/Cellar/python/3.7.7/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/robotide/spec/libraryfetcher.py", line 29, in get_import_result
) for kw in lib.handlers]
File "/usr/local/Cellar/python/3.7.7/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/robotide/spec/libraryfetcher.py", line 29, in
) for kw in lib.handlers]
File "/usr/local/Cellar/python/3.7.7/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/robotide/spec/libraryfetcher.py", line 40, in _parse_args
parsed[index] = parsed[index] + '=' + str(args.defaults[value]) # DEBUG str(value)

macos Catalina 10.15.5
RIDE v2.0b1 running on Python 3.7.7.
robotframework==3.2.1

There is need to provide alias parameter to Open Available Browser keyword.

User comments about the problem https://robocorp-developers.slack.com/archives/CH1GSKJ3U/p1596380462275000 onwards

workaround for the problem was to create custom keyword

*** Keywords ***
My RPA Browser
    [Arguments]   ${alias} 
    ${chrome_options}=  Evaluate  sys.modules['selenium.webdriver'].ChromeOptions()  sys, selenium.webdriver
    Call Method    ${chrome_options}    add_argument    --disable-extensions
    Call Method    ${chrome_options}    add_argument    --headless
    Call Method    ${chrome_options}    add_argument    --disable-gpu
    Call Method    ${chrome_options}    add_argument    --no-sandbox
    Create Webdriver    Chrome    chrome_options=${chrome_options}  alias=${alias}

The keyword should take a selector and use it to input text into an element, i.e. textbox or similar.

I'm following the docs with regards to RCC.

The following conda.yaml file is given as an example:

channels:
  - defaults
  - conda-forge
dependencies:
  - python=3.7.5
  - pip=20.1
  - nodejs=14.2.0
  - pip:
    - rpaframework==6.7.1
    - robotframework-browser==2.3.3
rccPostInstall:
  - rfbrowser init

However with v7.0 including Playwright, I had assumed the following would work:

channels:
  - defaults
  - conda-forge
dependencies:
  - python=3.7.5
  - pip=20.1
  - nodejs=14.2.0
  - pip:
    - rpaframework==7.0.*
rccPostInstall:
  - rfbrowser init

However this throws an error because it can't find 'rfbrowser`

===  new live  ---  post install phase ===
Running post install script 'rfbrowser init' ...
Script 'rfbrowser init' failure: Cannot find command: rfbrowser

I'm able to work around this by installing 'robotframework-browser' but then I'm still able to use RPA.Browser.Playwright.

channels:
  - defaults
  - conda-forge
dependencies:
  - python=3.7.5
  - pip=20.1
  - nodejs=14.2.0
  - pip:
    - rpaframework==7.0.*
    - robotframework-browser==2.3.*
rccPostInstall:
  - rfbrowser init

What's the preferred strategy to use v7 and the playwright post install process?

Thanks

As the title suggests, RPA Framework is quite big because of its dependencies and it would be better if it was composed of multiple packages. That would give users the benefit of having more granular control of runtime dependencies. A top-level metapackage could still be maintained for easier development workflows.

So I have an Excel file with some data in it and I am trying to know the first row.

from RPA.Excel.Application import Application
app = Application()
app.open_application(True)
app.open_workbook(r"C:\Users\User\Downloads\file.xlsx")
free_row = app.find_first_available_row("Sheet1")

And get this error:

Traceback (most recent call last):
File ".\rpa_test.py", line 12, in
tupla_libre = app.find_first_available_row("Sheet1")
File "C:\proyectos\rpa_pruebas.venv\lib\site-packages\RPA\Excel\Application.py", line 120, in find_first_available_row
cell_value = worksheet.Cells(row, column).Value
AttributeError: 'str' object has no attribute 'Cells'

I am using rpaframework==6.4.0
Windows 10 64bits

to be included in the RPA.Cloud.Google library

this also "forces" to implement support of OAuth2 authentication using credentials.json

keywords

• Init Drive Client
• Drive Upload Files
• Drive Download Files
• Drive Update File (actions=trash, untrash, star, unstar)
• Drive Delete File
• Drive Get Folder Id
• Drive Search Files
• Drive Move File
• Drive Create Directory

https://developers.google.com/drive

Hello , can you help please with this errror ?

The code I'm running is this:

from RPA.Browser import Browser
nav = Browser()
nav.open_available_browser("www.google.com", headless=True)

but print this error:

Traceback (most recent call last):
  File "robot.py", line 1, in <module>
    from RPA.Browser import Browser
  File "/home/andres/.local/lib/python3.8/site-packages/RPA/Browser.py", line 14, in <module>
    from robot.libraries.BuiltIn import BuiltIn, RobotNotRunningError
  File "/home/andres/dev/robot/robot.py", line 1, in <module>
    from RPA.Browser import Browser
ImportError: cannot import name 'Browser' from partially initialized module 'RPA.Browser' (most likely due to a circular import) (/home/andres/.local/lib/python3.8/site-packages/RPA/Browser.py)

There are various issues with the current PDF library implementation, in particular with finding elements and scraping data reliably. This issue is used for gathering all improvements ideas and requests to the libary.

Template matching and OCR should also work with websites through the webdriver, instead of having to use desktop automation for it. Integration to rpaframework-recognition should be quite easy, but reliable full-page screenshotting in selenium will require some thinking.

When automating applications with RPA.Desktop.Windows, it might give an errors such as: AttributeError: 'NoneType' object has no attribute 'wait'.

This seems to be a regression that happens when it fails to attach to a window. It should raise a proper error or fallback to a sane default (top window?) but instead continues normally. Other keywords do not check if the property is None and attempt to access the dialog, which results in the error message above.

Reported initially in Robocorp forum: https://forum.robocorp.com/t/windows-desktop-application-robot-tutorial/82/2

• rpaframework-core and rpaframework-recognition (and their RF dependencies) as conda-forge PR conda-forge/staged-recipes#13329
• rpaframework conda-forge/staged-recipes#13387

This is my configuration:

I updated it right after the release of rpaframework 2.5.1
I can work around the issue using keywords "Row Count" and "Should Be Equal As Integers" instead.

RPA.Tasks with a schema file is ignoring FAIL transitions: { "status": "FAIL", "task": "final" }

create object model of the responses and return that from approriate existing keywords

General information:

Windows 10 64bits
Python 3.8.6
rpaframework==6.5.0

Steps to reproduce:

from RPA.Excel.Application import Application

app = Application()
app.open_application(visible=True)
app.open_workbook(r'C:\pry\rpaframework\challenge.xlsx')
app.set_active_worksheet(sheetname='Sheet1')
last_row = app.find_first_available_row()

for x in range(last_row[0]):
row = app.read_from_cells(worksheet="Sheet1", row=x, column=1)
print("A: " + row)

Error:

Traceback (most recent call last):
File ".\test_rpa.py", line 23, in
row = app.read_from_cells(worksheet="Sheet1", row=x, column=1)
File "C:\proyectos\rpaframework.venv\lib\site-packages\RPA\Excel\Application.py", line 222, in read_from_cells
cell_value = worksheet.Cells(row, column).Value
AttributeError: 'str' object has no attribute 'Cells'

challenge.xlsx

UnicodeEncodeError: 'latin-1' codec can't encode characters in position 933-934: ordinal not in range(256).

Ran with latest Lab + RPAFW using work-item-to-pdf example.

two reports on this
(1) Win10 Finnish language, Chrome browser, Finnish language (above log is from this case)
(2) Chrome browser, Japanese language we are getting UnicodeDecodeError when trying to insert Japanese text into a web form using RPA.Browser. Is there any provisions for working around this error, or do we need to abandon Robocorp for another RPA tool?

More details about the (2) at https://robocorp-developers.slack.com/archives/CV9MR721E/p1595676282245500

Any future support for Java GUI? The Desktop.Windows library seems only works for Windows Form.

Using:
rpaframework==7.0.1
rpaframework-core==5.0.0

I would like use the list_worksheet() function, but the objects throws an error that it doesn't exist at all.

Upon checking the attributes of the object, it seems like the method list_worksheet() function doesn't exist at all on the Excel object.

Is there any way to retrieve the file owner info/file metadata with the RPA.FileSystem library?

Some keywords in the RPA.Browser always expect to receive strings, which creates surprising errors. For instance Select From List By Value does not function as expected when given an integer.

The library should have type hints for automatic RF argument conversion and/or explicit casts where certain types are expected.

RPA.Tables: Read Table From CSV splits a single column header by the letter 'a'.

For example schema_name becomes ['schem', '_n', 'me']

I'm working with the following code:

@{lst_a}= Create List a b c
@{lst_b}= Create List 1 2 3
&{dct_t}= Create Dictionary name=${lst_a} age=${lst_b}
${table_t}= Create Table ${dct_t}
Log ${table_t.index[0]}
&{_}= Pop Table Row table=${table_t}

Which results on:
ValueError: Unknown index name: Table(columns=['name', 'age'], rows=3)

This is raised by Table._location on line 337:
raise ValueError(f"Unknown {name} name: {value}")

The cause is rooted on an extra parameter when invoking get_row, on line 1021:
values = table.get_row(table, index, as_list)

It should be:
values = table.get_row(index, as_list)

Replacing "Pop Table Row" with this:
Set Column As Index table=${table_t} column=name

I get this other error:
TypeError: Keyword requires Table object

This is raised by Tables. requires_table on line 868:
raise TypeError("Keyword requires Table object")

The cause is rooted on a missing table parameter when invoking Tables.pop_table_column on line 1055:
values = self.pop_table_column(column)

It should be:
values = self.pop_table_column(table, column)

Currently, template matching and OCR are only supported in a desktop automation context, but a library (probably RPA.Images) should be made that supports the same functionality but with a user-given document instead.

Intermittently when running Firefox in headless mode, it will crash with the message: WebDriverException: Message: Failed to decode response from marionette.

Related upstream issue: mozilla/geckodriver#1559

Mobile automation libraries:

• Appium

Domain specific libraries:

• Jira
• Trello
• Word

Utility libraries:

• XML
• SSH
• LDAP, AD
• Crypto

?? libraries (Need more details):

• Java
• XML-RPC
• SOAP
• REST
• Network

WHY:

Robotframework-4.0a3 introduces support for IF/ELSE syntax which we see beneficial for
RPA and generic automation use case. Compared to test automation where Robot Framework has origins in,
RPA processes need to take various error scenarios into account and also the flow control in general may be complex. Branching operations with the current Robot Framework syntax make the code hard to read
and understand and for this IF / ELSE support is great improvement.

WHAT:

List of RF4 changes in 4.0b1
We probably need to support both RF 3 & 4..

• Backwards incompatible changes
  • Output.xml differences caused by removal of criticality and addition of SKIP
    • Probably affects executors, including jupyterlab robot kernels
    • could affect log analyzer
  • Argument conversion changes, could affect RPA framework libraries
  • Libdoc spec changes, require / enable changes at documentation generation tooling
    • Argument name, type and default are stored separately. (#3578)
    • Spec files have new information such as Enum and TypedDict data types. (#3607)
    • Argument name, type and default are stored separately. (#3578)
• Start the work with alpha / pre-release so that we are ready to go
• Document Lab and VsCode restrictions.
• Warn users about Robot features that might break

Work items:
[ ] RPA Framework supports for RF 4
[ ] Depended libraries support RF 4
[ ] Docs to support 3 & 4

Improvement idea from support channel:

1. For one side DataDriver applies a Keyword (called the "Test Template") to a collection of data, all running a test independently of each other. So if one fails, the rest are allowed to run and resolve on their own terms regardless.

2. On the other hand, you have RPA.Tasks, that allows to define an execution flow via a JSON file, referred to as the "schema".

They're both are incompatible to each other, given that both take ownership of the flow of tasks... and here's the suggestion:

Wouldn't it be nice to have something like RPA.Tasks that instead of applying to the Tasks section it applied to the subtasks of a keyword? Like the Template Keyword used by DataDriver... this way you could have tests running independently of each other and at the same time, define the execution flow via a schema.

Sometimes webdriver (at least chromedriver) fails to work with the following error:
FileNotFoundError: [Errno 2] No such file or directory: '/var/folders/rx/hx27566s54n5ws8gvpn7fd840000gn/T/drivers/chromedriver'

Removing the entire folder manually resolves the issue. This seems to be an upstream issue with webdrivermanager. Figure out if a patch can be made upstream or worked around here.

We should maybe also figure out a better place to store downloaded webdrivers.

It seems that some dependencies are lacking pre-built wheels for Python 3.9. Check which ones and upgrade as necessary.

Reported originally on forums: https://forum.robocorp.com/t/experimental-library-tasks/51/2

Currently the Robocloud.Items library is built around the convention of Robocloud work item payload being a JSON object with another variables object containing key/value pairs. This is a simple and sensible convention for many use cases.

However, especially with integrations to external systems, the payload format cannot always be controlled. One such example is launching a process on Typeform subscription, as discussed here: https://forum.robocorp.com/t/how-to-trigger-a-robot-from-typeform/100. Robocloud itself does not have constraints for the schema of the payload (apart from size limit).

In addition to the opinionated structure with variablesas key/value pairs, RPA framework could provide means to access the entire work item payload object.

The RPA.Desktop library should have support for OCR in Desktop automation.

It should support the following core features:

• The ability to locate text/numbers in screenshots, e.g. locate coordinates of a button based on text “Save”.
• The ability to read text/numbers from screenshots, e.g. read address information from textbox.

We already have Input Text When Element Is Visible. Could we also have Click Element When Visible? Almost all web robots need that one since more and more content is loaded dynamically.

For example:

Accept cookie consent
    ${accept_cookies_button}=    Set Variable    css:.browser-consent button.accept
    Wait Until Element Is Visible    ${accept_cookies_button}
    Click Button    ${accept_cookies_button}

Accept cookie consent
    Click Element When Visible    css:.browser-consent button.accept