haversine | Go library which implements the haversine formula

Check your output distances, what units are they? I converted mine to kilometers. You can check using an online distance calculator if you wanted. Let me know

Welcome to the site.

You are off to an "OK" start. Your model has quite a few errors in it.... did you look at the examples in the pyomo documentation?

A few suggestions:

1. Start with ConcreteModel and manually initialize the data. I think it is easier to do, esp. with python's ability to handle .csv files either manually (as I show below) or with pandas or csv_reader.

2. Throw out pandas for now. Use it from a separate file if need be to create the .csv files if you are comfortable with that, or just manually write them, or use csv_reader, but don't commingle pandas and pyomo till you get your feet on the ground. Same advice for numpy.

3. Use "flat file" format for your data, not tabular. It is easier to ingest. So, for example, create your distance table in a csv that has 3 columns as mine does and it is easier to read into a dictionary or, if you go to an AbstractModel it is in an easy format.

4. Use a small slice of your data and pprint() your model to make sure it makes sense and complies with your math model.

There are multiple problems:

1. If you want a Turkish model, you can try to find a pretrained Word2Vec model for Turkish (e.g. check out this repository) or train a model for Turkish yourself. The way you use it now you seem to train a model but only from a single website, which will barely do anything because the model needs a large amount of sentences to learn anything (like at least 10.000, better much more). Also you set min_count=5 anyway, so any word appearing less than 5 times is ignored generally. Try something like training it on the Turkish Wikipedia, see the linked repository.

2. Word2Vec by default is a unigram model, so the input is a single word. If you hand it a bigram consisting of two words like "Jaccard mesafesi" it will not find anything. Also you should catch the case that the word is not in vocabulary, otherwise each unknown word will cause an error and your program to cancel. Search for the unigram representation of each token, then combine the two, e.g. by using the statistical mean of the vectors:

To be faster, you should use compiled code.

pyproj allows to calculate a distance between 2 points: (Geod.line_length)

I think you misunderstand what DETERMINISTIC means as an option to CREATE FUNCTION.

It does not mean the result of the function is memoized. There is no cache of function results. There is no command to refresh this cache, because the results are not kept.

The meaning of DETERMINISTIC mainly affects binary logging:

https://mariadb.com/kb/en/create-function/#not-deterministic

The [NOT] DETERMINISTIC clause also affects binary logging, because the STATEMENT format can not be used to store or replicate non-deterministic statements.

That is, a non-deterministic function may return different results if it is executed on a replica, so if you use the function in an SQL statement that modifies data, the binary log format for that event must be ROW to ensure the same change is applied on the replica.

There's also a vague reference:

The optimizer may choose a faster execution plan if it known that the function is deterministic.

But no example is given for such a case. This is likely to be a rarity for it to make a significant performance difference.

I don't think this will be an effective performance optimization for your use case.

Both are used for two different purposes

You can use an apply with a lambda here to work on single rows. When you input df['start_coord'] you're using the whole series.

You cannot directly pass pd.Series to haversine function.

Instead of pulling the data straight up from Firestore and putting into the FutureBuilder, why don't you move this functionality inside your provided service UserProvider, for example (or another service), that way you can encapsulate your data fetching logic and your ordering logic there, as such: