Help me with Decision about increasing Application speed

Anvil Q&A
21

This mean my problems in my code :frowning:
I should now understand all suggestion, test them and give feedback to all

2 Likes
22

There is no free lunch that tastes better than taking something you built, doing some small changes and seeing a 1000X improvement in performance.

Optimization is like finish work for a project, you might think about it while you are building, but you still need to polish at the end to make it look really good and it can be very satisfying.

6 Likes
23

If you have specific forms that are slow, and can create a smaller clone that shows the performance problem, post a clone link here and youā€™ll get more specific advice about improving the performance.

3 Likes
24

Thank Ian for your suggestions !
I will try to implement
I have code on Server and Data on databases - Forms calls and uses Data
that cause low running speed
Now I am thinking How to organize code again with less anvil.servercalls

25

TIL (on Reddit) that thereā€™s a term for that strategy, or for a piece of it.

Edit: This is supported by Anvilā€™s Portable Classes. A DTO is likely to be composed of multiple values, of various types. All of them need to know how to [de]serialize themselves for transfer over the wire. With Anvilā€™s support, you can build custom DTOs (and subobjects) that know how to do just that.

3 Likes
26

It would be great to have a feedback about ALL round trips, not only about the ones we expect: Add logging of all round trips

1 Like
27

When I placed

print(datatime.daytime.now())

all over, it made me realize where the time goes There were usually simple fixes, as already mentioned in preceding posts.

4 Likes
28

Every generation of programmers rediscovers this for themselves: the only effective way to optimize begins with measurement.

Glad you stuck with it. :smiley:

2 Likes
29

A post was merged into an existing topic: Latency dependence on geographical location

30

Itā€™s been a while but if anybody has similar issues: -
I would introduce logging to your server functions with microseconds so you know what is taking a longtime, something like this: -

def write_log_msg(str_func_name, str_state, str_type, str_message):
  '''
  str_function   string   calling this function
  str_state   string   state calling function is in i.e. start or executing or exiting
  str_type   string   INFO or DEBUG or WARN or EXCEPTION 
  str_message   string   only required by the state executing; the text of the log message to be logged
  e.g. puts log line like '(05-05-2023,09:29:17.513506),(anvil username),function_name,INFO,Entered' 
  to the Anvil log 
  '''
  obj_current_time = datetime.now()
  if str_state == 'entry':
    print(f'{obj_current_time.strftime("%d-%m-%Y,%H:%M:%S.%f,(%z)")},{anvil.users.get_user()},{str_func_name},INFO,Entered')
    return
  elif str_state == 'exit':
    print(f'{obj_current_time.strftime("%d-%m-%Y,%H:%M:%S.%f,(%z)")},{anvil.users.get_user()},{str_func_name},INFO,Exiting')
    return
  elif str_state == 'executing':
    print(f'{obj_current_time.strftime("%d-%m-%Y,%H:%M:%S.%f,(%z)")},{anvil.users.get_user()},{str_func_name},{str_type},{str_message}')
    return
  else:
    print(f'{obj_current_time.strftime("%d-%m-%Y,%H:%M:%S.%f,(%z)")},{anvil.users.get_user()},{write_log_msg.__name__},DEBUG,Incorrectly Called str_func_name=({str_func_name}), str_state=({str_state}), str_type=({str_type}), str_message=({str_message})')
    return

then whenever I enter, leave (a function) or face an exception I record the time, for example:-

def get_db_conn():
  ''' 
  creates a connection (object) to MySQL DB 
  str_db_host   string   global variable is the hostname
  int_port  integer   global v ariable for the DB listener port
  returns   obj_dbconn   a pymysql connection object
  '''
  write_log_msg(str_func_name=get_db_conn.__name__, str_state=entry)
  try:
    obj_dbconn = pymysql.connect(host = str_db_host,
                               port = int_port,
                               user = anvil.secrets.get_secret('db_user'),
                               password = anvil.secrets.get_secret('db_password'),                               
                               cursorclass = pymysql.cursors.DictCursor,
                               #ssl={"fake_flag_to_enable_tls":True})
                               #ssl={"ssl_verify_cert":True},
                               ssl={"ssl_verify_identity":True})
    
  except pymysql.err.OperationalError as error:
    # an exception creates a 4 item list that can contain info it would be pymysql.err.args[] 0to4
    write_log_msg(str_func_name=get_db_conn.__name__, str_state='executing', str_type='DEBUG', str_message='Got MySQL Error Code=({error.args[0]}), MySQL Error Text=({error.args[1]})')
    write_log_msg(str_func_name=get_db_conn.__name__, str_state='exit')
    exit()
  else:
    write_log_msg(str_func_name=get_db_conn.__name__, str_state='exit')
    return obj_dbconn

This will give you a starting point to see which functions are slow and then you can introduce further logging to figure which part is particularly slowā€¦

The code didnā€™t quite format well but it will give someone something to hack to their likingā€¦

EDIT
edited to format the code correctly

2 Likes
31

For readability, you can (and should) wrap your code as shown below:

``` python
your code here
```

Without that, indents and some punctuation will be lost.

1 Like
32

Thanks, updated the postā€¦

33

So much clearer for everyone. Thank you!

34

If you are really into programming and performance. Get the Python libraries to do the computation as they are compiled C code. For example matrices and arrays use Numpy rather than writing Python code/lists to do it.

https://www.youtube.com/watch?v=tD5NrevFtbU
Balance clean code vs. performance

OOP is not always the best approach. I only use OOP if I see it has objects, structured programming, implement data-orientated programming concepts
https://www.youtube.com/watch?v=rlrDefjpNqY&t=194s
You use Python data-classes instead of C/C++ structs

https://www.youtube.com/watch?v=08CWw_VD45w&t=370s
functional programming and how we can use these ideas in Python: -
https://www.youtube.com/watch?v=4B24vYj_vaI

2 Likes