google sheets – Conditional Formatting w/multiple conditions and partial text match

cells that do not match the conditions are getting highlighted and cells with the conditions are being omitted

That sound like you need a NOT to negate the result.

column C’s matching row begins with “EM”

The regular expression now matches all content which contains EM, so you probably need to replace the argument with "^EM"; the caret indicates the beginning of the content. (That’s a standard regex operator; as indicated here, Google Sheets uses the RE2 version.)

You could try this:


and see whether it works for you.

forms – Standard style for text that opens hover tool tip?

Most applications/websites either have an icon (often a question mark) to indicate you can hover over it (but then the hover only works on the icon, not the label) or no special formatting at all. The answer here suggests using a dotted line as well; I vaguely remember old Windows (3.x) help files working the same way.

Stack Exchange has a lot of labels / links with tooltips:

  • almost all dates/timestamps (the ‘asked 14 mins ago’, which appears on your user card while I’m writing this, the two ‘today’s currently under the question title)
  • the links under each post (‘share’, ‘edit’, ‘follow’, etc.)
  • most cases where a number larger than 1000 isn’t shown exactly, such as the number of views of a question

They do not have a common style. Perhaps if they would, users on a touchscreen device would wonder about what the common style would mean; after all, they are unable to access hovers.

html – Localization of multi style text element in Vue

After following up this question, I’m thinking if there’s a better/shorter way to write this text element that has part with different style, while allowing localizations:

    <span class="has-text-grey">{{ $store.releaseList != null ? $store.releaseList.length : 0 }}</span>
    <span> versions released</span>

With vue-i18n:

<p v-html="$t('downloads.releases.table.versions-total')
    .replace('{0}', '<span class=has-text-grey>{1}</span>')
    .replace('{1}', $store.releaseList != null ? $store.releaseList.length : 0)">

The localizable text would be {0} versions released.

I could set the localizable text to <span class=has-text-grey>{0}</span> versions released, but that would make the life of the translators (usually people without knowledge about HTML) a bit more difficult.

Is there any better built in way to add support for localizations in a multi-styled text?

I will create a 30-45 second introductory text animation for $15

I will create a 30-45 second introductory text animation

I am a self-taught text animator. As a product or service provider, experimenting with videos is a sure way of delivering mobile-friendly content to your customers. 

I provide animation for ads, explainer videos and commercials. I can help you create short videos for your website, Insta feed or twitter. 

I will write your script for an additional price.


Please contact me to find out more details.


machine learning – How to extract an open research question from text with Natural Language Processing?

Researchers often broadly state one or several research problems as an “open research question” or an “open research problem”, a “research gap” or “desideratum” or often make “suggestions for further research”. The same terms are used across disciplines. Some disciplines are referring to categories of research gaps like an “evidence gap” (commonly used in medical research). There are probably less than a dozen terms used to describe the same thing.

As long as you are dealing with quantitative research (from natural to social sciences) these open research questions are mostly asked or at least repeated in the “conclusion”, “future research” or “discussion” section of an article. There are a very limited number of phrases used to introduce a desideratum like “further research… (is needed to/will show etc.)…”, “it remains to be seen, if…”. One could prepare a quite comprehensive list of such phrases rather quickly. Often these phrases also signify the beginning of the statement, which should be helpful.

In other words, you know what keywords and phrases to look for and have a measure of relevance by there position in the document. Also many open research problems are directly phrased as a question.

There has to be an element of human curation (aided by further NLP analysis), since it will be difficult to extract the exact statement and the statement has to be validated as non-redundant and preferably categorized (if applicable as “evidence gap”, “method gap” “sample gap” etc.)

Assuming you have a database of full-text articles (most pre-print servers give full access to content and meta-data), others give you free access to metadata and scientometric data. What methods, algorithms and software solutions (open source or SaaS) would you use or test?

API’s of science repositories and the access they give:

sharepoint online – extract “all” numerics from text string

I have a Sharepoint Online list with a text field for a contact called Mobile Phone…
I want to extract the numerics only…

(111) 222-3333 would be 1112223333
+99 111-222-3333 would be 991112223333
111-222-3333 would be 1112223333
111 222 3333 would be 1112223333

I was expecting to be able to do this with a calculated column, but it’s not as obvious as I thought. Ultimately our dialing system needs a numerics only version of the value (no special characters or spacing). I have seen similar questions answered here but not this exact need. Thanks in advance…

encryption – Can original text be extracted from data that has gone through HMACSHA512, if both salt and hash is available?

Can original text be extracted from data that has gone through HMACSHA512, if both salt and hash is available ?

I think, believe(and hope) that the answer is NO, but since I’m not a security expert in anyway, I would like to have some reaffirmation from people who do know the intricacies.

Recently there have been quite a few data leaks happening around(5+ in past 2 months), due to which the users personal details are getting exposed to the internet. So, it makes me wonder about a few things. Is it a good idea to store the hashed form of all user identity related data(eg. email, phone, dob, etc) instead of having them stored as plain text? What would be the obvious pros and cons(except performance, of course) of doing this? Is doing this worth it ?

PS: Please do not down vote, I’m only trying to get a better understanding and clarification. Thanks.