Postgres performance

[max-hassani]

The new class is intended to monitor important parameters of pyiron central database. Here, it is assumed that the database is a Postgresql database. The final data is presented as pandas dataframe.

the use case should be :

from pyiron_base import DatabaseStatistics
db_stat = DatabaseStatistics()
df = db_stat.performance()

The references for the queries can be found:
[1] https://russ.garrett.co.uk/2015/10/02/postgres-monitoring-cheatsheet/
[2] https://pgdash.io/blog/essential-postgres-monitoring-part2.html
[3] https://wiki.postgresql.org/wiki/Index_Maintenance

[max-hassani]

Up to an extent, I converted the sql queries to sqlalechemy expressions, but some of them are really long and complicated to convert. But if the reviewers see it essential, I will try to convert all the queries to sqlalchemy expressions.

[max-hassani]

Of course a unittest will be added. I have to first figure out how to create a postgres database via a github workflow.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[jan-janssen]

I find this code hard to read. From my perspective it would be easier to return the entry rather than writing to the dictionary. This also simplifies debugging later.

[max-hassani]

Do you mean the whole code or this function in particular? There are multiple functions which modifies the self._performance_dict. Some of them, like here and here would be more complex to be returned an entry. But definitely can be done.

[jan-janssen]

Yes, I would like each function to return a sub dictionary and then use dict.update() to merge them in the end. https://python-reference.readthedocs.io/en/latest/docs/dict/update.html This way we can debug the individual functions, look at the output they return. Following the zen of python: Explicit is better than implicit. https://www.python.org/dev/peps/pep-0020/

[max-hassani]

Thanks for the links. Now each method returns a dictionary which is used at the end to update self._performance_dict.

[jan-janssen]

@muh-hassani I added a few comments but apart from those I guess it would be great to merge this functionality soon.

[jan-janssen]

I like the class still I can not stop wondering if it would not be easier for the user to have just one function?

from pyiron_base import get_database_statistics
get_database_statistics()

This function could be defined as:

def get_database_statistics():
    return DatabaseStatistics().performance()

[max-hassani]

I like the class still I can not stop wondering if it would not be easier for the user to have just one function?

from pyiron_base import get_database_statistics
get_database_statistics()

This function could be defined as:

def get_database_statistics():
    return DatabaseStatistics().performance()

I agree, it makes the usage much easier. I have done it, the way you suggested.
I also made two methods as separate functions, as they did not need to be part of the class. You can find them here and here

[max-hassani]

@jan-janssen, if you approve the current status of this PR, I can merge it.

[jan-janssen]

@jan-janssen, if you approve the current status of this PR, I can merge it.

Looks good to me.