Winnipeg | Handling Missing Data in R with MICE

Put simply: your activate() call in form_valid() comes too late to affect the form field, so the incoming datetime gets interpreted in the default timezone—which in your case is America/Toronto—before being converted to UTC and saved to the database. Hence the apparent time shift.

The documentation doesn't really specify when you need to call activate(). Presumably, though, it has to come before Django converts the string value in the request to the aware Python datetime in the form dictionary (or vice versa when sending a datetime). By the time form_valid() is called, the dictionary of field values is already populated with the Python datetime object.

The most common place to put activate() is in middleware (as in this example from the documentation), since that ensures that it comes before any view processing. Alternatively, if using generic class-based views like you are, you could put it in dispatch().

There's a thing called tz, zoneinfo, or the Olson database. It's maintained by the Internet Assigned Numbers Authority

It names zones in Continent/City or sometimes Continent/State/City format, like Asia/Kolkata or America/Indiana/Knox. Each named zone contains rules for converting to and from UTC time to local time, including the correct handling of summer time.

The database contains the temporopolitical history of time zone and summer time changeovers for the city (and surrounding regions). So, if the government of, say, Knox Indiana USA, changes the summer time rules next year, their entry gets updated in a maintenance release of the database.

If Spain decides to repudiate its Franco-era decision to use the same time zone as 'Europe/Berlin', and move to the same zone as 'Europe/Lisbon', the updated zoneinfo data for 'Europe/Madrid' will reflect that change on its effective date. Updates to UNIX-based operating systems like Linux and MacOS include the most recent zoneinfo data.

php uses zoneinfo. So does MySql. moment.js adds synonyms to it. If somebody in Knox sets their time zone to moment's synonym 'US/Central' and the city council the changes the rules, they won't follow the change.

So, please consider using php's list in your application, because it's proven so far to be future-proof.

^\S+(?=\s\S+$)

This regex gives you the first word of all teamnames that only consist of two words. The others you have to sort manually, because there is no way to tell just by pattern if the middle word is part of the city or the teamname.

You were almost right in your code. You had to do out = out.append({...}, index=False) for the dict that you wanted to append. There was some logic mistake in your code. I fixed it below (along with some fixed for readibility):

Possible and Probable issue: Multithreading and assignment to Instance level reference.

TLDR; in AddPushpin() , use local variable instead of instance variable. replace this.Pin = new Pushpin() with var pin=new Pushpin() and then add this local variable to collection (You might want to change this collection with threadsafe one)

Explanation:

In pandas, the agg operation takes single or multiple individual methods to be applied to relevant columns and returns a summary of the outputs. In python, lists hold and parse multiple entities. In this case, I pass a list of functions into the aggregator. In your case, you were parsing a dictionary, which means you had to handle each column individually making it very manual. Happy to explain further if not clear

How difficult this is depends on what info you want from the page. I am working to the assumption you want a dataframe/tibble that details from PointOfCare e.g. 1. Hospitals, through ServiceType e.g. Publicly Funded / Free Services, all the way down into the actual listings details of each service.

There are two immediate problems to overcome if going for all the above info:

1. The DOM is pretty flat i.e. the PointOfCare info is at same level of DOM as ServiceType and the start of service listings is only 1 level deeper. This means there is no nice logical way to use an HTML parser and select for parent nodes then process children, and still get the desired info mapped for the PointOfCare and ServiceType to each service listing.
2. There are differing numbers of child nodes holding a given service's info, those with className summaryRecordType, within each listing (ranging between 3 and 5).

① To deal with the first problem I decide to convert the retrieved HTML to a string and split that string into chunks to process. I retrieve the PointOfCare labels and use those to generate the initial blocks settings_blocks:

Since you seem to only want to match exact words regex is way more expensive then just looking the words up. Assuming you only need to match whole words and not a complicated regular expression (e.g. numbers etc.) you can split the description into words and perform a lookup. If the words are saved in sets lookup will be O(1)

Code would look something like this

Would this work? You only have to rename the city column once at the end.