Ecosyste.ms: Awesome

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

Awesome Lists | Featured Topics | Projects

https://github.com/wisehackermonkey/searchfusion

makes it super easy to find stuff on your computer
https://github.com/wisehackermonkey/searchfusion

Last synced: 1 day ago
JSON representation

makes it super easy to find stuff on your computer

Awesome Lists containing this project

README 

        
# Search fusion 



### makes it super easy to find stuff on your computer

## local gui app that allows you to search for images and text files by tags autogenrated by ml classifiers



# install

# install conda

```bash

cd ~

mkdir -p ~/miniconda3

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh

bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3

rm -rf ~/miniconda3/miniconda.sh

~/miniconda3/bin/conda init bash

```

# close and open terminal

```bash

conda --version

conda create -n searchfusion python=3.8

conda activate searchfusion

pip install 'transformers[torch]' pillow

```



# examples of just using the classifer script

```bash

python blip.py --text "A " --input_file ./images/muffin.png --verbose > classified_images.txt

 ```



### Explanation

1. **Argument Parsing**: Added `--verbose` flag to enable verbose mode.

2. **Logging Configuration**: Configured logging to suppress all messages except critical ones by default. If `--verbose` is enabled, it shows warning messages.

3. **Logging Messages**: Added `logging.warning` calls at key points to provide messages when verbose mode is enabled.



### How to Run

1. **Using default values and suppressing messages**:

   ```bash

   python script.py < image_file.png

   ```



2. **Using verbose mode**:

   ```bash

   python script.py --verbose < image_file.png

   ```



3. **Providing a text prompt, input file, and verbose mode**:

   ```bash

   python script.py --text "A cute " --input_file ./images/muffin.png --verbose

   ```



This setup will suppress all messages except for the `print` output unless the `--verbose` flag is used.



## glob a bunch of files

cat ./images/crypto.png | python blip.py --text "A " >> classified_images.txt

cat ./images/crypto.png | python blip.py --text "photo of " --verbose >> classified_images.txt

ls ./images/*.* | parallel -j 2 'cat {} | python blip.py --text "a " --verbose >> classified_images.txt'



# todo add a hash value as a key

# add squel lite? idk

# interface?



# rip grep the file with fuzzy search



# search bar



# external resources

https://huggingface.co/Salesforce/blip-image-captioning-large



# target output

![alt text](image.png)



```bash

sudo apt-get install ripgrep

pip install ripgrepy

```



### prompt

```python

write a python script 



create a database using sqlite 

witht the following columns 



image_descriptions 

    path TEXT,

    description TEXT,

    hash TEXT

path (text) (path to file) (has the abspath of the image file)

description(text)

hash (text) (md5 hash of the file) make entry unique on this key



create the database call it image_descriptions



up_sert the file  "/home/o/github/searchfusion/images/tor.png" into the database

description = "a photo of a shady guy holding a onion"

db name "demo.db"

code only

```



# prompt

```python

create a python script 

that searches a sqlite database.

   db table name: image_descriptions

    path TEXT PRIMARY KEY,

    description TEXT,

    hash TEXT



key word search on description



create a commandline search input display the results 

for each result

", 



add command line arguments 

--db 'demo.db'

--query -q 'onion'

script name is search.py

code only, example use

```



# prompt

```python

create a python script 

file name: jobs.py

search all sub directories (recusivly) for images using find

input is a list of directories to search within example ['./images', './secret']



images = 'png, jpg, gif, webp, bmp'



create a priority queue store in a sqlite database called jobs.db table name jobs_queue

        path TEXT, (abspath)

        time (time created)

        hash TEXT UNIQUE 

        completed BOOL

create a sqlite database. db file name images.db

   db table name: image_descriptions

    path TEXT,

    description TEXT,

    hash TEXT UNIQUE PRIMARY KEY

NUMWORKERS = 2

if there is jobs in the queue the pull off NUM_WORKERS

and run the shell command  

parallel NUMORKERS 'python blip.py --text "photo of " --verbose  --input_file {path} '

if NUMWORKERS is more than 1 then run multiple  python blip.py --text "photo of " --verbose  --input_file {path} '



store the results of each job into the image_descriptions table



code only, example use

```



# prompt



```python

convert this python script to accept list of image files from ls ./path/to/dir/*.png | vit.py 

vit.py (name of the script)

example path = ./images/*.png 



intalize  processor and model at the top of the file 

```