Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/infinitifall/soft-string-compare

C++ functions to judge the similarity of strings. Originally created to correct messy human input. Also compiles to Web Assembly.
https://github.com/infinitifall/soft-string-compare

cmake cpp data-cleaning string-matching string-similarity wasm

Last synced: 4 days ago
JSON representation

C++ functions to judge the similarity of strings. Originally created to correct messy human input. Also compiles to Web Assembly.

Awesome Lists containing this project

README 

        
# Soft String Compare



C++ functions to judge the similarity of strings. Originally created to correct messy human input. Also compiles to WebAssembly.



## Install



You need `cmake` and a C++ compiler (such as `g++`) installed.



```bash

# clone repo

git clone https://github.com/Infinitifall/Soft-String-Compare

cd Soft-String-Compare



# make

cd build_native

cmake ../ -DCMAKE_BUILD_TYPE=Release

# cmake ../ -DCMAKE_BUILD_TYPE=Debug

cmake --build ./

```



## Use



The code in [example.cpp](./example.cpp) will try to correct the messy product names in [input_list.txt](data_dummy/input_list.txt) (see below) by matching them to a catalog of products.



```

iphne 13 pro maks

nkie ar jrdan 1

samsng 55" qled 4k smrt tv

dysin v11 absloote

instnt pot dou plus 6 qt

playstaton 5 digitl editon

fitbt versa 3 smrt wach

keurigk-cup coffe makr

bose quetcomfort 45 hdphnes

nutribullt pro - 13-pce hi-sped blndr

roomba i3+ evo self-emtying robt vacum

ninja foodie 10-in1 xl pro ar fyer & othr

lgo star wars milenim falcn



...

```



Run the `example` executable



```bash

# run

cd build_native

./example

```



```

✅ iPhone 13 Pro Max

✅ Nike Air Jordan 1

✅ Samsung 55" QLED 4K Smart TV

✅ Dyson V11 Absolute

✅ Instant Pot Duo Plus 6 Qt

✅ PlayStation 5 Digital Edition

✅ Fitbit Versa 3 Smartwatch

✅ Keurig K-Cup Coffee Maker

✅ Bose QuietComfort 45 Headphones

✅ NutriBullet Pro - 13-Piece High-Speed Blender

✅ iRobot Roomba i3+ EVO Self-Emptying Robot Vacuum

✅ Ninja Foodi 10-in-1 XL Pro Air Fryer & Other

✅ LEGO Star Wars Millennium Falcon

✅ Amazon Echo Dot (4th Gen) Smart Speaker

✅ VIZIO 5.1 Surround Sound Bar System

✅ Logitech MX Master 3S Wireless Mouse

✅ KitchenAid Artisan Series 5 Qt. Mixer

✅ GoPro HERO11 Black

✅ Apple Watch Series 7 GPS + Cellular

✅ Sonos One SL Wi-Fi Speaker

✅ Microsoft Surface Pro 8 Laptop

✅ LG 65" C1 Series OLED 4K UHD Smart TV

✅ Breville The Barista Express Espresso Machine

✅ Garmin Forerunner 945 GPS Running Watch

✅ Whirlpool 4.8 cu. ft. Front Load Washer

✅ Canon EOS R6 Mirrorless Camera

✅ Beats by Dr. Dre Studio3 Wireless Over-Ear Headphones

✅ Theragun Prime Deep Tissue Massage Gun

✅ Philips Norelco Multigroom All-in-One Trimmer

✅ Nespresso Vertuo Next Coffee & Espresso Maker

✅ Ring Video Doorbell Pro 2

✅ Brita Longlast Water Filter Pitcher

✅ Vitamix E310 Explorian Series Blender

✅ Traeger Pro 575 Wood Pellet Grill

✅ Oculus Quest 2 Advanced All-in-One Virtual Reality Headset

✅ Sunbeam Osmo 3 Reverse Osmosis Water Filter System

✅ Merax 10' Trampoline with Enclosure

✅ Klipsch HT-G700 3.1ch Dolby Atmos Soundbar

✅ YETI Tundra 45 Hard Cooler

✅ RTIC UltraLight 52 Qt Cooler

✅ HidrateSpark 3.0 32oz Insulated Water Bottle

❌ Bose QuietComfort 45 Headphones (lumin ultra-comfortble coper-infusd matres = Leesa Original Mattress)

✅ Anova Culinary Sous Vide Precision Cooker

✅ AeroGarden Harvest Elite Indoor Garden

✅ Waterpik Aquarius Water Flosser

✅ eufy RoboVac 11S (Slim) Robot Vacuum

✅ SKIL 1/4" Hex Electric Screwdriver

✅ SMOK Novo 4 Pod System Vape Kit

✅ Anker PowerCore 10000 Portable Charger



✅ count = 48

☑️  count = 0

❌ count = 1



🎯 ratio = 97.959 %

```



We see it is able to match severely misspelled product names to a very high degree (`>90%`).



We can also enter arbitrary strings to see how the system ranks items. For example, to see why `lumin ultra-comfortble coper-infusd matres` wasn't correctly matched:



```

Enter name: lumin ultra-comfortble coper-infusd matres

```



```

...



0.000 rating: Logitech MX Master 3S Wireless Mouse

0.000 rating: VIZIO 5.1 Surround Sound Bar System

0.360 rating: Vitamix E310 Explorian Series Blender

0.360 rating: NutriBullet Pro - 13-Piece High-Speed Blender

0.847 rating: Merax 10' Trampoline with Enclosure

0.847 rating: Philips Norelco Multigroom All-in-One Trimmer

2.357 rating: Anker PowerCore 10000 Portable Charger

2.659 rating: Nespresso Vertuo Next Coffee & Espresso Maker

3.701 rating: RTIC UltraLight 52 Qt Cooler

3.924 rating: Traeger Pro 575 Wood Pellet Grill

5.997 rating: Breville The Barista Express Espresso Machine

7.131 rating: Leesa Original Mattress

9.128 rating: Garmin Forerunner 945 GPS Running Watch

14.926 rating: Bose QuietComfort 45 Headphones



1: lumin ultra-comfortble coper-infusd matres

++ ____________comfort_______________________

++ ________________________________________es

++ _______________________co_________________

++ _____________________e ___________________

-- ___e __________________________

-- __________Co___________________

-- _____________________________es

-- __________Comfort______________

2: Bose QuietComfort 45 Headphones

```



We see that the correct choice `Leesa Original Mattress` is given the third highest rating. This is a particularly tricky example because it doesn't have much in common with the input string `lumin ultra-comfortble coper-infusd matres`.



## Performance



In [example.cpp](./example.cpp): In function `int main(int argc, char** argv)`: Comment out all lines except `real_run_wrapper(argc, argv);`. Compile with `Release` to enable optimization flags



```bash

# alternatively, run the `cmake release` task in VS Codium



cd build_native

cmake ../ -DCMAKE_BUILD_TYPE=Release

cmake --build ./

```



```bash

# using bash

time ./example ../data_dummy/all_list.txt ../data_dummy/input_list.txt ../data_dummy/output_list.txt ../data_dummy/all_list.txt

```



```bash

real    0m0.059s

user    0m0.052s

sys     0m0.003s

```



To correct 50 product names on a Dell XPS 15 (9510, 2021) equivalent.



## Compile to WebAssembly



You need a Wasm compiler (like `Emscripten`) and a local development server (such as `http-server`)



```bash

# make

cd build_wasm

emcmake cmake ../ -DCMAKE_BUILD_TYPE=Release

# emcmake cmake ../ -DCMAKE_BUILD_TYPE=Debug

emmake make



# run a local development server

npx http-server ./ -o -p 9999

# visit http://localhost:9999 in your web browser

```



Useful guides on C/C++ to Wasm:

- https://developer.mozilla.org/en-US/docs/WebAssembly/C_to_Wasm

- https://marcoselvatici.github.io/WASM_tutorial/



## Programming in VS Codium



- `clangd` extension for completions, references, errors and hints

- `CodeLLDB` extension works great with the provided [launch.json](./.vscode/launch.json) and [tasks.json](./.vscode/tasks.json)