Skip to content

Instantly share code, notes, and snippets.

@FLamparski
FLamparski / pm_sensor_wifi_rec.ino
Last active June 22, 2019 15:46
A program for collecting data from an SDS-011 PM 2.5 sensor
#include <ArduinoJson.h>
#include <WiFi.h>
#include <SdsDustSensor.h>
#include <PubSubClient.h>
// v1.0 - 2019-06-16 - Initial public release of script
// v1.1 - 2019-06-17 - Turn off sensor if resetting due to lack of connectivity
// v1.2 - 2019-06-22 - Increase WiFi connection timeout
// XXX Change these to your WiFi details
/**
* Computes the checksum of a SDS011 message.
*
* This checksum is the low 8 bits of the sum of the data bytes.
* The sensor will not react to commands with invalid checksum.
*/
byte sds011_checksum(byte* msg, size_t msg_size) {
int sum = 0;
for (int i = 2; i < msg_size - 2; i++) {
sum += msg[i];
@FLamparski
FLamparski / pm-basic.ino
Created June 5, 2019 10:48
Simple ESP32/Arduino program for reading the SDS-011 output in continuous mode
void setup() {
Serial.begin(115200);
Serial2.begin(9600, SERIAL_8N1, -1, -1, false);
}
void loop() {
if (Serial2.available()) {
byte pm_buf[10];
Serial2.readBytes(pm_buf, 10);
@FLamparski
FLamparski / simple-sds011.py
Created June 5, 2019 10:22
Simplest SDS-011 Python program
import serial
import struct
from datetime import datetime
# Change this to the right port - /dev/tty* on Linux and Mac and COM* on Windows
PORT = 'COM5'
UNPACK_PAT = '<ccHHHcc'
with serial.Serial(PORT, 9600, bytesize=8, parity='N', stopbits=1) as ser:
@FLamparski
FLamparski / plot.py
Created June 5, 2019 10:21
Example programs for the SDS-011
import serial
import struct
from collections import deque
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation as anim
from matplotlib.ticker import FuncFormatter
# Change this to the right port - /dev/tty* on Linux and Mac and COM* on Windows
PORT = 'COM5'
@FLamparski
FLamparski / README.md
Last active May 28, 2019 17:40
A JS program for playing with different, but related, search algorithms

This is an extension of the previous gist - while that gist explored a particular question, this one tries to compare some algorithms which have very similar implementations, but different requirements and performance characteristics. It generates a random but somewhat believable geography on a plane and then tries to get from Q to W (noting that there may be more than one connection between two cities, with different costs). Each run is different.

An example result:

====================================================================================
Algorithm A*:
Expanded Q; Agenda = [(E, 308), (P, 190), (D, 289), (A, 497)]
Expanded E; Agenda = [(P, 190), (W, 412), (D, 289), (I, 512), (A, 497)]
Expanded P; Agenda = [(W, 412), (D, 289), (L, 338), (M, 385), (I, 512), (G, 420), (A, 497), (V, 528)]
@FLamparski
FLamparski / README.md
Last active May 28, 2019 13:44
A JS program that replicates the mark scheme solution to the Uniform Cost Search question on ECS629 Artificial Intelligence past exam from 2015

If you run agendaSearch.js you should get the same output as output.txt

I reverse-engineered the optimisations made in the mark scheme solution. They were:

  1. If a cheaper path exists to a node in the agenda already, don't add the more expensive path to the agenda. If you find a cheaper path to a node than you already have in the agenda, remove the more expensive path.
  2. Don't expand nodes which were already expanded, even if you reach them through a different path. In UCS, the path to every node expanded so far should already be optimal.

The algorithm works even without those two optimisations, but the agenda size gets quite big - not an issue for a computer on such a small problem, possibly a problem for a computer on a much larger problem, definitely a problem for someone doing this in an exam.

To turn off the simplifying assumptions:

<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8"/>
<title>foreach or for(i)</title>
<script src="https://cdnjs.cloudflare.com/ajax/libs/benchmark/1.0.0/benchmark.min.js"></script>
<script src="./suite.js"></script>
</head>
<body>
<h1>Open the console to view the results</h1>
@FLamparski
FLamparski / notebook.ipynb
Last active July 20, 2018 16:48
Further adventures in reinforcement learning
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8"/>
<title>maps maps maps</title>
<script src="https://cdnjs.cloudflare.com/ajax/libs/benchmark/1.0.0/benchmark.min.js"></script>
<script src="./suite.js"></script>
</head>
<body>
<h1>Open the console to view the results</h1>