---
tags: comp-decomp
title: Meeting-notes
---

# Overview
Log meeting notes

## logs

### Day 1:
Had meeting with collaborator.
Keep the following in mind:

### Day 2:
- catching up with the code I was working with..
- working in `compare_score.py`
- command: `source curr_command.sh`

Working on the function `find_matching_structure` in file `cfg.py`. The aim of the function is to:
- create a string of edges.
- The edges are traversed in bfs manner
- The current function finds the largest matching string
- The string contains all edges as: `last_node_instr:first_node_instr`
- Limitation:
- matches only one subsequence
- need multiple matching
- Can match only incoming

Work:
- Added funciton `get_edges_of_node` to extract outgoing edges of a node in terms of string
- Will use this to find the CFG at node level
- Added `replace_common_strings` function to replace the recurring opcode in string with standard notations

### Day 3:
Using command: `python bin/compare_score.py test/initial.pyc \<module\> test/original.pyc foo
`

Goal: "Fetching lines from the binary code"

Work done:
- created `get_source_line` function in CFG class.
- created `get_line_no` function in CFGNode class.

Testing:
- Fetching line numbers and testing whether that works
- command: `python bin/test.py` -> Passing right now

Next goal:
- start perturbation to generate an output file with code.
- Get lines from the database
- Then add outgoing edges
- and then start perturbing

### Day 4:
Using command: `python bin/perturb_manually.py`

Added feature to manually perturb the following files:
- `test/temp.py`: Will be the file that you perturb manually
- `test/initial.pyc`: Will be the file that you generate as target

Compare the two (`test/temp.png` and `test/initial.png`) cfg to see the whether you are getting closer or farther away.

### Day X:

- Completed implementation of environment for manual perturbation
- We use the partial matching because reason here

---

# Meeting Hamza:

- thought of them
- Long term:
- Scoring algorithm optimized
- Meta-data
- cover samples and optimize
- Short term:
- Optimize seed utilization
- useful hints from seed