Wipy 2.0 and save data on my pc, not in flash memory

[VFlorio94]

@robert-hh
yes, the problem is the function 'calcolo_position(1)'. But after I call this function and I acquire data during 1 second, in the next call of the function the data collected before are not important.
So, I think that I need to refresh/ release this memory allocation after the call to the function but I don't know how.
Can you help me? thank you a lot.

[VFlorio94]

@robert-hh The first time the program terminates without error.
if I run for the second time the same main program don't go on and give me the message 'memory allocation...'

[VFlorio94]

@robert-hh I suppose that the problem can be the function 'calcolo_position', but i'm not sure, beacuse the message is only which I copy before.
the first time the program terminates in the right way, with no error and everything it's ok.
then I run the program newly and it don't work.
In my opinion, for my experience with c++ for example , I have to search to release the memory allocation and restart with the allocation from the begging each time that i call the function 'calcolo_position(1). but i don't know how. can you help me?
thank you for your help

[robert-hh]

@vflorio94 said in Wipy 2.0 and save data on my pc, not in flash memory:

hen I run the program newly and it don't work.

How do you restart the program?
Normally, when objects are not used any more, they are released and garbage collection will reclaim the space. You can fore garbage collection with the call to gc.collect().

[VFlorio94]

@robert-hh to restart the program I mantain the program open and I reuse the command run.
I put gc.collect() in the code, you can see where in the code that I copy to you before, but the error stay.

[robert-hh]

@vflorio94 Can you try to start and run the program from a simple terminal emulator like putty or screen. Using the intgrated environments of Atom or VSC may hide or create problems.

[VFlorio94]

@robert-hh Is IDLE Python GUI good?
now I need to implement code with more lines, to compare the other data that calcolo_position(1) return in the same function 'acquisizione_interval_pitch()'.
I'm afraid about this, because I think that the error can be show before the end of the acquisition.

And more I have to use an interval of acquisition more longer than 20 second, like 120 second.

[robert-hh]

@vflorio94 You can use any editor you like for writing the code on you PC. The PyCom plugin for upload and run is what may create unexpected result. To sort that out, upload your code with ftp (Filezilla) and use a simple serial terminal to run the code.

[VFlorio94]

@robert-hh yes, I'll try to use GUI Idle python , but I have some problem with socket

[VFlorio94]

@robert-hh I have some difficult to use a serial terminal to run the code, because I have 'import machine' in main, so I have to use socket, but I don't understand how.
Can you hel me?

#Sending
import socket

DEST_IP = "192.168.4.255"
UDP_PORT = 15005
sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)

def sending(msg):
    sock.sendto(msg, (DEST_IP, UDP_PORT))

#main
import machine 
import pycom
import time
import lib_imu
import gc
import interval_position
import sending
#rilevazione posizione centrale corretta
print("INIZIO ACQUISIONE")
print("Rilevazione posizione corretta")
time.sleep(1)
position_ok = []
position_ok = interval_position.calcolo_position(6, 1)
mean_pos_ok = position_ok[0]
print("media posizione corretta", mean_pos_ok[2])
dev_std_ok = position_ok[1]
gc.collect()
print("FINE Rilevazione posizione corretta")
time.sleep(2)
print("Rilevazione posizione errata")
time.sleep(1)
position_no =[]
position_no = interval_position.calcolo_position(6, 2)
mean_pos_no = position_no[0]
dev_std_no = position_no[1]
gc.collect()
print("FINE Rilevazione posizione errata")
time.sleep(1)
print("Inizio test pitch")
time.sleep(1)
# the function below return 1 if the test is completed correctly
p = interval_position.acquisizione_interval_pitch(20, 1,  mean_pos_ok, mean_pos_no, dev_std_no, 3)
del position_ok
del position_no
if p == 1:
    stringa = "Acquisizione terminata"
    sending.sending(stringa)

[robert-hh]

@vflorio94 Using a serial terminal is not at all in conflict with "import machine". What you might see is, that once you runa program on your device, the REPL prompt is not available any more.

[VFlorio94]

@robert-hh yes , when I run rhe code in Visual Studio code with this main I have this output :

Calcolo media
Acquisizione terminata
fine prova
╝╝>

the program end correctly. And the REPL prompt is not avaiable any more.
Now i don't have anymore the message of error with memory allocation.
but now I try to complete the code with more lines and more implementation

[VFlorio94]

@robert-hh And I 'm not able to use serial terminal

[VFlorio94]

@robert-hh Sorry, but I need your help one more if is possible.
I have to acquire data from IMU for a long time ( 20 minuts) and write the data on file.
but also this time I have the error

╝Traceback (most recent call last):
 File "<stdin>", line 22, in <module>
MemoryError: memory allocation failed, allocating 2048 bytes
╝>

Here there are the simple code lines that I wrote. This error is terrible for me, probably I make an error on how to do this . please , can you Help me please?

import machine 
import pycom
from machine import Timer
import time
import lib_imu
from machine import I2C
from machine import PWM

chrono = Timer.Chrono()
lib_imu.set_configmode()
time.sleep(0.07)
lib_imu.set_units()
time.sleep(0.3)
lib_imu.set_mode('ndof')
List =[]
chrono.start()
while 1 :
    lap = chrono.read()
    if lap <= 1200: 
        eul = lib_imu.get_euler()    
        List.append(eul[0])
        time.sleep(0.02)
    else:
        print("Salvataggio su file")
        file_dati = open("dati360.txt", "w")
        file_dati.write(repr(List[1:len(List)]))
        file_dati.close()
        print("fine")
        chrono.stop()
        del List
        break

[robert-hh]

@vflorio94 So you want first to collect for 1200 seconds data, with a delay of 20 ms each. I do not know how long taking a value needs, if have someting in mind like ~100 ms, which give a list of 10000 entries, and then you want to store that. AFAIK, the basic mp object type needs 8 bytes/entry, so you need 80kbytes for the fill table. Being already low on memory, the list append fails soon, when the runtime engine of micropython wants to allocate more memory for the list, roughly after 128 cycles.
So no, you simply cannot do that with a WiPY2. You have to write to SD in shorter cycles, even if that would interrupt your data acquisition.
Alternatively, use a device with more memory, like a WiPy3.

[VFlorio94]

@robert-hh So, if i want to collect as more possible data in 1200 s how frequently can i take values?

[robert-hh]

@vflorio94 You should collect & write at the same time. Since the writikng is buffered, most of the write operations will not take long.

[robert-hh]

@vflorio94 The code could look like:

import machine 
import pycom
from machine import Timer
import time
import lib_imu
from machine import I2C
from machine import PWM

chrono = Timer.Chrono()
lib_imu.set_configmode()
time.sleep(0.07)
lib_imu.set_units()
time.sleep(0.3)
lib_imu.set_mode('ndof')
with open("dati360.txt", "w") as file_dati:
    chrono.start()
    lap = chrono.read()
    while lap < 1200
        file_dati.write(repr(lib_imu.get_euler()[0]))
        time.sleep(0.02)
        lap = chrono.read()
print("fine")
chrono.stop()

[VFlorio94]

@robert-hh Thank you, I did in this way and the code run, but the real sample isn't 0.02 s but each 0.035 s. And I don't know why?!

[robert-hh]

@vflorio94 Taking the measurement and writing takes time too. Now you know, it's about 0.015s.