GY-63_MS5611/libraries/TOPMAX
2023-06-17 18:01:24 +02:00
..
.github 0.1.0 TOPMAX 2023-05-18 21:21:08 +02:00
examples 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
test 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
.arduino-ci.yml 0.1.0 TOPMAX 2023-05-18 21:21:08 +02:00
CHANGELOG.md 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
keywords.txt 0.1.1 TOPMAX 2023-06-13 19:16:10 +02:00
library.json 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
library.properties 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
LICENSE 0.1.0 TOPMAX 2023-05-18 21:21:08 +02:00
README.md 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
TOPMAX.cpp 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00
TOPMAX.h 0.2.0 TOPMAX 2023-06-17 18:01:24 +02:00

Arduino CI Arduino-lint JSON check License: MIT GitHub release

TOPMAX

Arduino library to track the top N maxima.

Description

This experimental library tracks the top N maxima of a series of values.

The library implements two classes:

  • TOPMAX just tracks the values. (smaller footprint).
  • TOPMAXext also tracks a user defined 32-bit tag (e.g. timestamp).

The latter is more useful as it allows to link the top N values to a certain time or day, an index in a database, or another indirect reference. It can even be used to store a second parameter e.g. what was the humidity for the top N temperatures.

The library can be used for different applications:

  • track top N peak temperature / humidity over a given period of time.
  • track top N UV peak intensity during the day.
  • track top N trading volumes or .
  • track top N largest fish caught in a contest.
  • track top N number of visitors during a day.
  • track top N of time-outs, e.g. to monitor production loss.

The 32 bit tag, can be filled with a 32 bit number like a timestamp, bit can also be used more creatively e.g. 2x 16 bit numbers or even 32 booleans. Any mapping is possible (but not part of the library).

0.2.0 breaking change

Since version 0.2.0 the maxima is found with getValue(index == 0). In earlier versions the maxima was found with index == count().

The advantage of this change is that independent of the size of the TOPMAX object the true maxima will always at index == 0. So if you change your code and the size of the TOPMAX object, far less code needs to be changed. It also allows to have two or more TOPMAX objects of different size and use one index to access both.

Interface

#include "TOPMAX.h"

TOPMAX

  • TOPMAX(uint8_t size = 5) Constructor, defines the number of elements it can hold. Default number of elements is 5. If size < 3 it will be set to 3. The maximum size is currently 255, or less if there is not enough memory.
  • uint8_t count() returns the number of elements in the internal array. 0.. size.
  • uint8_t size() returns the maximum number of elements in the internal array.
  • void reset() reset the internal counter to 0, logical clearing of the system.
  • bool add(float value) add a value to the TOPMAX object if it is in the top N of maxima. If so the smallest element is removed. Returns false if not added or if there was an allocation error.
  • float getValue(uint8_t index) get an element of the internal array. The index must be <= count(), if not the function currently returns NaN. This may or may not be a valid value, so the user should guard the index parameter carefully.
  • bool fill(float value) convenience function to fill the internal array with a single value e.g. 0. Returns true on success.

TOPMAXext

Derived from TOPMAX, extended with a tag field.

  • TOPMAXext(uint8_t size = 5) Constructor, defines the number of elements it can hold. Default number of elements is 5. If size < 3 it will be set to 3. The maximum size is currently 255, or less if there is not enough memory.
  • bool add(float value, uint32_t tag) add a value to the TOPMAXext object to check if it needs to be in the top N of maxima. If so add the 32-bit tag too (at same index). The 32-bit tag is typical an index, counter or timestamp, but any semantic is possible.
  • uint32_t getTag(uint8_t index) get the tag from an element of the internal array. The index must be <= count(), if not the function currently returns 0xFFFFFFFF. This may or may not be a valid value, so the user should guard the index parameter carefully.
  • bool fill(float value, uint32_t tag) convenience function to fill the internal array with a single value e.g. 0. (tag idem).

Future

Must

  • keep TOPMIN and TOPMAX functional in sync.
  • improve documentation.

Should

  • add unit tests.
    • for extended version if possible.

Could

  • add more examples.
    • example creative use of tag field.
  • add error handling
    • TOP_ERR_ALLOCATION
    • TOP_ERR_INDEX
    • TOP_NOT_ADDED
    • TOP_OK
    • int error; int lastError()
  • create template class.
  • bool check(float value) and bool check(float value, uint32_t tag)
    • if (value < getValue(0)) or so.
    • inRange(value)
    • checks if this value would be added to the TOP-N
  • TOP class, in which the condition can be set as parameter.
    • a function() returning true or false when comparing 2 values.
  • bool hasValue(float value) and bool hasTag(uint32_t tag)
    • or int getIndex(...) duplicates?

Wont

  • how to handle double/triple etc. entries with same value
    • they are handled as unique elements, that is what I had in mind.
  • optimize loops in fill() (maybe upon request)
    • pointer math