add raspberry pi and other fast-gpio device support when no pulseio

[ladyada]

tested on Pi 3

[jerryneedell]

I ran this on a Raspberry Pi 3B+
It works - sometimes ...
I get bad readings quite often (every 3 or 4) but there is an occasional crash -- see example:

>>> import dht_simpletest
Temp: 85.1 F / 29.5 C    Humidity: 68.3% 
A full buffer was not returned.  Try again.
Temp: 85.1 F / 29.5 C    Humidity: 68.7% 
Checksum did not validate. Try again.
Temp: 84.9 F / 29.4 C    Humidity: 68.9% 
Temp: 84.9 F / 29.4 C    Humidity: 68.9% 
Temp: 84.9 F / 29.4 C    Humidity: 68.9% 
A full buffer was not returned.  Try again.
Temp: 84.9 F / 29.4 C    Humidity: 69.0% 
Checksum did not validate. Try again.
Temp: 84.7 F / 29.3 C    Humidity: 69.0% 
Checksum did not validate. Try again.
Temp: 84.7 F / 29.3 C    Humidity: 69.0% 
Temp: 84.7 F / 29.3 C    Humidity: 69.0% 
Temp: 84.7 F / 29.3 C    Humidity: 68.9% 
Checksum did not validate. Try again.
Temp: 84.7 F / 29.3 C    Humidity: 68.9% 
Temp: 84.7 F / 29.3 C    Humidity: 68.7% 
Temp: 84.7 F / 29.3 C    Humidity: 68.6% 
A full buffer was not returned.  Try again.
Temp: 84.6 F / 29.2 C    Humidity: 68.5% 
Temp: 84.6 F / 29.2 C    Humidity: 68.5% 
A full buffer was not returned.  Try again.
Temp: 84.6 F / 29.2 C    Humidity: 68.4% 
A full buffer was not returned.  Try again.
Temp: 84.6 F / 29.2 C    Humidity: 68.4% 
Temp: 84.4 F / 29.1 C    Humidity: 68.4% 
Temp: 84.4 F / 29.1 C    Humidity: 68.5% 
Temp: 84.4 F / 29.1 C    Humidity: 68.5% 
Temp: 84.4 F / 29.1 C    Humidity: 68.6% 
Temp: 84.4 F / 29.1 C    Humidity: 68.6% 
Temp: 84.4 F / 29.1 C    Humidity: 68.6% 
Temp: 84.4 F / 29.1 C    Humidity: 68.6% 
Temp: 84.4 F / 29.1 C    Humidity: 68.6% 
A full buffer was not returned.  Try again.
Checksum did not validate. Try again.
Checksum did not validate. Try again.
Temp: 84.2 F / 29.0 C    Humidity: 68.5% 
Temp: 84.2 F / 29.0 C    Humidity: 68.6% 
Temp: 84.2 F / 29.0 C    Humidity: 68.6% 
Temp: 84.2 F / 29.0 C    Humidity: 68.6% 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/pi/projects/blinka/dht/dht_simpletest.py", line 11, in <module>
    temperature_c = dhtDevice.temperature
  File "/home/pi/projects/blinka/dht/adafruit_dht.py", line 213, in temperature
    self.measure()
  File "/home/pi/projects/blinka/dht/adafruit_dht.py", line 164, in measure
    pulses = self._get_pulses()
  File "/home/pi/projects/blinka/dht/adafruit_dht.py", line 149, in _get_pulses
    pulses.append(int(1000000 * (transitions[i] - transitions[i-1])))
OverflowError: unsigned short is greater than maximum
>>> 

[jerryneedell]

I made this change and it seems to run OK: Lots of invalid checksums and "Full buffer not returned " errors, but it keeps running without the integer overflows.

diff --git a/adafruit_dht.py b/adafruit_dht.py
index f795f4d..928c50f 100644
--- a/adafruit_dht.py
+++ b/adafruit_dht.py
@@ -146,7 +146,11 @@ class DHTBase:
                         transitions.append(time.monotonic()) # save the timestamp
                 # convert transtions to microsecond delta pulses:
                 for i in range(1, len(transitions)):
-                    pulses.append(int(1000000 * (transitions[i] - transitions[i-1])))
+                    pulses_micro_sec = int(1000000 * (transitions[i] - transitions[i-1]))
+                    if pulses_micro_sec > 65535:
+                        pulse.append(65535)
+                    else:
+                        pulses.append(pulses_micro_sec)
         return pulses
 
     def measure(self):

[ladyada]

awesome, added!

[tannewt]

How much more memory does this take up in a SAMD? (You can measure with gc.mem_free() with a gc.collect before it.) Thanks!

[jerryneedell]

Is there a reason why this has not been merged yet? It just came up via the forum:
https://forums.adafruit.com/viewtopic.php?f=60&t=142347

[ladyada]

umm i didnt get around to calculating the gc.free on an m0 before and after
@tannewt do ya still want that?

[brennen]

i can check mem usage here in a minute.

[tannewt]

@brennen let us know what you find.

[brennen]

ok, so i was running the following test on a feather m0 express:

import time
from board import D6
import gc

gc.collect()
print("before import:", gc.mem_free())
import adafruit_dht
print("after import:", gc.mem_free())

gc.collect()
print("before init:", gc.mem_free())
#initial the dht device
dhtDevice = adafruit_dht.DHT22(D6)
print("after init:", gc.mem_free())

while True:
    gc.collect()
    print("before sensor read:", gc.mem_free())
    try:
        time.sleep(.5)
        # show the values to the serial port
        temperature = dhtDevice.temperature * (9 / 5) + 32
        humidity = dhtDevice.humidity
        print("Temp: {:.1f} F Humidity: {}% ".format(temperature, humidity))
        print("after sensor read:", gc.mem_free())
    except (RuntimeError, TypeError) as error:
        print(error.args)

on master, i get:

before import: 18912
after import: 11024
before init: 16464
after init: 16400
before sensor read: 16400
Temp: 69.6 F Humidity: 42.0% 
after sensor read: 15808
before sensor read: 16384

all well and good. with changes in this PR, i get:

main.py output:
before import: 18912
after import: 10320
before init: 15984
after init: 15920
before sensor read: 15920
('DHT sensor not found, check wiring',)
before sensor read: 15808
("unsupported types for __mul__: 'NoneType', 'float'",)
before sensor read: 15760
("unsupported types for __mul__: 'NoneType', 'float'",)
before sensor read: 15760
("unsupported types for __mul__: 'NoneType', 'float'",)
before sensor read: 15760
('DHT sensor not found, check wiring',)

increasing time.sleep() call to 2 seconds gets rid of the TypeError - i assume temperature was just returning None, which may be correct behavior that the simpletest just needs to be tweaked to account for, but it still seems like i can't get a reading at all with this code. here's the simpletest output:

main.py output:
DHT sensor not found, check wiring
DHT sensor not found, check wiring
DHT sensor not found, check wiring
DHT sensor not found, check wiring
DHT sensor not found, check wiring

[brennen]

...aaaaand i see i probably got my wires crossed with another merge to master. @jerryneedell lemme know if you want a memory check once conflicts are resolved.

[jerryneedell]

Hmm, I don’t the sensor not found error is not the same as the None issues we saw. I’ll try to reproduce this tomorrow.

[jerryneedell]

hmm -- I reproduced the errors reported by @brennen on a Trinket M0
I also found that there was a slight change made to the get_pulses() function in the PR.
If I revert it back to the previous code, I no longer get errors and the DHt sensor reads.
here is the change:

jerryneedell@Ubuntu-Macmini:~/projects/trinket_m0$ diff adafruit_dht.py ~/projects/test_DHT/Adafruit_CircuitPython_DHT/adafruit_dht.py
122,124c122,123
<                 while True:
<                     if time.monotonic() - tmono > 0.25:
<                         break # time out after 1/4 seconds
---
>                 while time.monotonic() - tmono > 0.25:
>                     pass # time out after 1/4 seconds

In the PR code , while loop will immediately end - the test is inverted - it should wait at least .25 sec.
this change also works and just fixes the relational test

jerryneedell@Ubuntu-Macmini:~/projects/trinket_m0$ diff adafruit_dht.py ~/projects/test_DHT/Adafruit_CircuitPython_DHT/adafruit_dht.py
122,123c122,123
<                 while time.monotonic() - tmono < 0.25:
<                         pass # time out after 1/4 seconds
---
>                 while time.monotonic() - tmono > 0.25:
>                     pass # time out after 1/4 seconds
jerryneedell@Ubuntu-Macmini:~/projects/trinket_m0$ 

Here is the output of @brennen test code on the trinket M0:

Adafruit CircuitPython 4.0.0-alpha.1-99-g293345119 on 2018-10-12; Adafruit Trinket M0 with samd21e18
>>> import dht_brennen
before import: 19472
after import: 16784
before init: 16784
after init: 16704
before sensor read: 16704
Temp: 73.2 F Humidity: 59.5%
after sensor read: 16112
before sensor read: 16688
Temp: 73.2 F Humidity: 59.5%
after sensor read: 16608
before sensor read: 16688
Temp: 73.2 F Humidity: 59.5%
after sensor read: 16608
before sensor read: 16688
Temp: 73.2 F Humidity: 59.5%
after sensor read: 16608
before sensor read: 16688
Temp: 73.2 F Humidity: 59.2%
after sensor read: 16112
before sensor read: 16688
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "dht_brennen.py", line 27, in <module>
  File "dht_brennen.py", line 20, in <module>
KeyboardInterrupt:

I'm not sure how to try to make this change to the PR and resolve the conflicts with the current master.
Can I do it with the Web Editor? @ladyada - would you prefer to just do the changes yourself.
Let me know how you want to proceed.

[ladyada]

sorry i have no idea how to fix and im not going to have time this week to learn, the changes arent that large.

i can just close this, y'all cna start a new PR?

[jerryneedell]

@ladyada I'll work on w new PR - hopefully tomorrow to fix this and reconcile with the master. Once it is in, you can close this one if you approve.

[ladyada]

yay thanks!

[jerryneedell]

@ladyada since #18 has been merged should we go ahead and close this?