Change in README

README.md

@@ -6,7 +6,7 @@
 This is a CLI tool to detect cuts in films, especially old films with noisy and broken frames. This tool basically takes an input video and stores cuts in various formats - frame index of cuts alongwith type of cut in CSV format, timestamp in seconds of start and end shots in the film, MEP json format containing shots timestamps, and a format which is supported by cinemetrics for further analysis. 
 This tool was produced for Google Summer Code 2021 with RedHenLabs and Media Ecology Project. 
 
-Detailed working of this tool is included in the [final submission blog](https://gsocblog-tre3x.netlify.app/final-project.html).
+Detailed working of this tool is included in the [final submission blog](https://tre3x.github.io/final).
 
 ## Getting Started
 
@@ -17,7 +17,8 @@ You need Python 3.x and Conda package manager to run this tool
 ### Installation
 For installing this tool with pretrained model, follow the steps below :
 1. Clone this repository `git clone https://github.com/tre3x/FilmEditDetection.git`
-2. Install the neccessary dependencies by executing `conda env create -f environment.yml`
+2. Switch to `cutdetector-CLI` branch `git checkout cutdetector-cli`
+3. Install the neccessary dependencies by executing `conda env create -f environment.yml`
 
 ### Dataset Description
 Two datasets were used in training the validation/softcut-detection module.
@@ -91,16 +92,6 @@ module load singularity/3.7.1
 ```
 6. Run the image
 ```bash
-singularity run -B <path/to/video> -B <path/to/model> filmedit.img --vidpath <path/to/video> --modpath <path/to/model> 
+singularity run FilmEditDetection.img --vidpath <path/to/video>
 ```
 - `<path/to/video>` denotes the absolute input video path
-- `<path/to/model>` denotes the absolute path to trained model. If empty, the tool will tun on pre trained model.
-
-## Future Work
-- Additional features can be added to the tool which provides helpful insights into the film metrics.
-- There are few cases in which the tool fails to classify cuts, like in case of Jump cuts. In the future some other variant of the model
-can be used to classify the cuts which the current model fails to classify.
-- Some films have low contrast with high noise, in which it becomes difficult to classify cuts. Some preprocessing on training data to acheive the same
-and let the model learn those features.
-- The command line interface can currently handle one output operation at a time. It can be modified 
- to allow the user to generate outputs of multiple format at a time.

main.py

@@ -14,7 +14,7 @@ def run_tool(vidpath, modpath, operation, outdir, config, cinemetrics_submit, yn
         if outdir=='':
             outdir = os.path.join(here, "csv_shots")
         run(conf).run(vidpath, modpath, outdir, iscsvtime=True)
-    if operation=='mepformat':
+    if operation=='mepjson':
         if outdir=='':
             outdir = os.path.join(here, "json_mep")
         run(conf).run(vidpath, modpath, outdir, ismepjson=True)