EC552 HW1

Description

Tool for modifying Cello Input file for improved genetic circuit performance

Authors

• Anirudh Watturkar
• Liam Murray

Inputs:

1. .v : Input Verilog file, representing an circuit.
2. .UCF.json : A JSON representation of the Library used by Cello, representing the chassis organism that the genetic circuit will exist within.
3. .input.json : A JSON representation of the Library used by Cello, representing the input signals into the genetic circuit will exist within.
4. .output.json : A JSON representation of the Library used by Cello, representing the output of the genetic circuit.

Outputs:

1. Score : The new score of the circuit (in the command line)
2. Delta : Difference between the new and old scores of the circuit (in the command line)
3. .NEWinput.json : The modified input file.

Installation Instructions

1. Ensure python 3.8+ & pip are installed

2. Install Docker

3. Pull docker instance

docker pull cidarlab/cello-dnacompiler:latest

4. Install requirements via pip:

pip install -r requirements.txt

• Libs used:
  • numpy v1.20.2
  • pandas v1.2.3
  • celloapi2 v0.2.4 (has the following dependencies):
    • python-dateutil v2.8.1
    • pytz v2021.1
    • PyYAML v5.4.1
    • six v1.15.0

5. Ensure cello docker instance is running

6. Create File Structure for running Cello

There are two ways to do this:

1. Clone the homework 1 template, then place cli.py, optimizeInput.py, and evaluate.py in the top-level folder of the repository. Then, create an output directory.

   git clone https://github.com/CIDARLAB/homework1-template

   Your file structure should now be like this:

   ├── input
   │   ├── and.v
   │   ├── Eco1C1G1T1.input.json
   │   ├── Eco1C1G1T1.output.json
   │   ├── Eco1C1G1T1.UCF.json
   │   ├── nand.v
   │   ├── options.csv
   │   ├── struct.v
   │   └── xor.v
   ├── cli.py            <-
   ├── evaluate.py       <-
   ├── main.py           x
   ├── optimizeInput.py  <-
   ├── output
   ├── poetry.lock       x
   ├── pyproject.toml    x
   └── README.md         x
   

2. Recreate the above file structure and populate the input directory with the input, output ucf, options.csv, and verilog files. Note: the files marked with 'x' next to them in the above structure do not need to be recreated if you are doing this manually.

Running

There are 3 ways to perform optimization & evaulate the results:

1. Using cli.py

Navigate to the directory where cli.py is found, then follow the instructions below:

-> To load *input.json, optimize its paramters, and save to a new file:

python cli.py -m <og_input_file.json> <new_file_name.json> <chassis_name>

ex: 
$ python cli.py -m Eco1C1G1T1.input.json Eco1C1G1T1NEW.input.json Eco1C1G1T1

-> To run Cello with a given *input.json file:

python cli.py -r <input.json> <num_input_signals> <chassis_name> <verilog_file.v>

ex:
$ python cli.py -r Eco1C1G1T1NEW.input.json 2 Eco1C1G1T1 and.v

-> To compare the original input to the optimized one:

python cli.py -mr <og_input_file.json> <new_file_name.json> <num_input_signals> <chassis_name> <verilog_file.v>

ex:
$ python cli.py -mr Eco1C1G1T1.input.json Eco1C1G1T1NEW.input.json 2 Eco1C1G1T1 and.v

-> For full usage instructions:

$ python cli.py -h

2. Using evaluate.py

If you cloned the template repo, then you should be able to simply run

$ python evaluate.py

For custom input files to Cello make the following changes to evaluate.py:

-> If you are using a chassis other than Eco1C1G1T1, modify the chassis_name variable in evaluate.py

-> If the circuit requires other than 2 inputs, set signal_input equal to the number of inputs to the circuit

-> If a custom verilog file is desired, set v_file to the new verilog file name

-> If a different options file is used, changed options to the new options file name

Notes:

• Make sure that there are folders named output and input, which contains the 4 input files listed above in the Inputs section.
• Detailed output returned from Cello will be stored in the output folder.
• The output score and delta values will be printed to the command line; the .NEWinput.json file will save to the input file.

3. With your own custom script

If you have your own custom script and would like to integrate our optimization script into it, place optimizeInput.py in the same directory as your script, then add the following to the top of your script:

from optimizeInput import *

Then, to get the optimal input file via our optimization method, add the following in your script (replacing the filenames as necessary):

input_class_list = get_input_models_in_class('Eco1C1G1T1.input.json')
gate_class_list = get_file_gate_models_in_class('Eco1C1G1T1.UCF.json')
mod_inputs_list = compute_optimal_parameters(input_class_list, gate_class_list)
save_input_class_in_file(mod_inputs_list,'Eco1C1G1T1.input.json','Eco1C1G1T1.NEWinput.json' )