Samar

.gitattributes

@@ -0,0 +1 @@
+*.ipynb filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -1,7 +1,20 @@
+# Ignore large files and folders
+models/
+data/
+output/
+*.pt
+*.pth
+*.ckpt
+# Ignore data, checkpoints, and models
+data/
+models/
+output/
+*.pt
 models/checkpoints/*
 datasets/data/*
 __pycache__/
 output/
 
 .venv/
-.conda/
+.conda/*.h5
+*.h5

README.md

@@ -1,6 +1,6 @@
-# L-CRP-TEMA
+# PCX-TEMA
 
-This repository contains the code for applying the L-CRP method for TEMA project. 
+This repository contains the code for applying the PCX method for TEMA project. 
 
 ## Setting up
 
@@ -10,17 +10,16 @@ Data and models are available on Google Drive https://drive.google.com/drive/fol
 
 Path for data is `datasets/data`.
 
-#### 1. U-Net model: 
-- Checkpoint: Checkpoints/unet_flood.pt
-- Dataset: Data/General_Flood_v3.zip
-- Task: specifically for this model - flood detection
+#### 1.PIDNet model: 
+- Checkpoint: Checkpoints/flood_model.pt
+- Dataset: Data/flood_segmentation.zip
+- Task: specifically for this model - flood segmentation
 
 #### 2. YOLOv6s6 model:
 - Checkpoint: Checkpoints/best_v6s6_ckpt.pt
 - Dataset: Data/PersonCarDetectionData 
 - Task: person and car detection
 
-#### 3. PIDNet model: not yet available
 
 
 ### Build the Docker image
@@ -124,4 +123,28 @@ python test_post_data.py ImageMetadata --cloud
 The test script will send sample image metadata to the application and you should see the processing results in the logs of both the Flask application and the worker process.
 
 
+### References
+
+
+LRP (Layer-wise Relevance Propagation):
+
+- original paper: (https://doi.org/10.1371/journal.pone.0130140)
+- overview paper: (https://doi.org/10.1007/978-3-030-28954-6_10)
+- zennit toolbox: (https://github.com/chr5tphr/zennit)
+
+CRP (Concept Relevance Propagation):
+
+- paper: (https://doi.org/10.1038/s42256-023-00711-8)
+- zennit-crp toolbox: (https://github.com/rachtibat/zennit-crp)
+
+
+L-CRP (Concept Relevance Propagation for Localization Models):
+
+- paper: (https://arxiv.org/pdf/2211.11426)
+- L-CRP code: (https://github.com/maxdreyer/L-CRP/tree/main)
+
+
+PCX (Prototypical Concept-based Explanations):
+- paper: (https://arxiv.org/pdf/2311.16681)
+- PCX code: (https://github.com/maxdreyer/pcx/tree/main)
 

app.py

@@ -15,8 +15,8 @@
 # Import your existing modules
 from src.explanator import Explanator
 from src.minio_client import MinIOClient, FHHI_MINIO_BUCKET, NAPLES_MINIO_BUCKET
-from common_app_funcs import update_entity, get_bm_id, set_bm_id, update_job_status, get_job_status, get_redis_conn, get_job_queue
-from tasks import process_image_task, explanator
+from common_app_funcs import update_entity, get_bm_id, set_bm_id,get_uav_id,set_uav_id,get_flight_number,set_flight_number, set_alert_ref_id , get_alert_ref_id , update_job_status, get_job_status, get_redis_conn, get_job_queue
+from tasks import process_image_task
 
 
 # Set up Redis connection and queue
@@ -232,9 +232,15 @@ def post_data():
 
         # Quick validation check
         if entity_type == "Alert":
+            uav_id = entity["uav_id"]["value"]
+            flight_number = entity["flight_number"]["value"]
             bm_id = entity["bm_id"]["value"]
-            set_bm_id(bm_id)
-            msg = f"Received Alert with bm_id and saved to redis: {bm_id}"
+            alert_ref = entity["alert_ref"]["object"]
+            set_uav_id(redis_conn, uav_id)
+            set_flight_number(redis_conn, flight_number)
+            set_bm_id(redis_conn, bm_id)
+            set_alert_ref_id(redis_conn, alert_ref)
+            msg = f"Received Alert with bm_id and alert_ref and uav_id and flight_number saved to redis: {bm_id}, {alert_ref}, uav_id: {uav_id}, flight_number: {flight_number}"
             return jsonify({'message': msg}), 200
 
         explanator = get_explanator()
@@ -245,18 +251,61 @@ def post_data():
             return jsonify({'error': err_msg}), 400
 
         current_bm_id = get_bm_id(redis_conn)
+        current_alert_ref_id = get_alert_ref_id(redis_conn)
+        current_uav_id = get_uav_id(redis_conn)
+        current_flight_number = get_flight_number(redis_conn)   
+        app.logger.debug(f"Current alert_ref: {current_alert_ref_id}")
         app.logger.debug(f"Current bm_id: {current_bm_id}")
+        app.logger.debug(f"Current uav_id: {current_uav_id}")
+        app.logger.debug(f"Current flight_number: {current_flight_number}")
 
         # Extract image information
-        posted_bm_id = entity["bm_id"]["value"]
-        if posted_bm_id != current_bm_id:
-            app.logger.warning(f"Received bm_id: {posted_bm_id} does not match current bm_id: {current_bm_id}")
+        posted_uav_id = entity["uav_id"]["value"]
+        if current_uav_id is None:
+            app.logger.info(f"No cached uav_id; storing value {posted_uav_id}.")
+            set_uav_id(redis_conn, posted_uav_id)
+            current_uav_id = posted_uav_id
+        elif posted_uav_id != current_uav_id:
+            app.logger.warning(f"Received uav_id: {posted_uav_id} does not match current uav_id: {current_uav_id}; updating stored uav_id.")
+            set_uav_id(redis_conn, posted_uav_id)
+            current_uav_id = posted_uav_id
+
+        posted_flight_number = entity["flight_number"]["value"]
+        if current_flight_number is None:
+            app.logger.info(f"No cached flight_number; storing value {posted_flight_number}.")
+            set_flight_number(redis_conn, posted_flight_number)
+            current_flight_number = posted_flight_number
+        elif posted_flight_number != current_flight_number:
+            app.logger.warning(f"Received flight_number: {posted_flight_number} does not match current flight_number: {current_flight_number}; updating stored flight_number.")
+            set_flight_number(redis_conn, posted_flight_number)
+            current_flight_number = posted_flight_number
+
+
 
+        posted_bm_id = entity["bm_id"]["value"]
+        if current_bm_id is None:
+            app.logger.info(f"No cached bm_id; storing value {posted_bm_id}.")
+            set_bm_id(redis_conn, posted_bm_id)
+            current_bm_id = posted_bm_id
+        elif posted_bm_id != current_bm_id:
+            app.logger.warning(f"Received bm_id: {posted_bm_id} does not match current bm_id: {current_bm_id}; updating stored bm_id.")
+            set_bm_id(redis_conn, posted_bm_id)
+            current_bm_id = posted_bm_id
+
+        posted_alert_ref = entity["alert_ref"]["object"]
+        if current_alert_ref_id is None:
+            app.logger.info(f"No cached alert_ref; storing value {posted_alert_ref}.")
+            set_alert_ref_id(redis_conn, posted_alert_ref)
+        elif posted_alert_ref != current_alert_ref_id:
+            app.logger.warning(f"Received alert_ref: {posted_alert_ref} does not match current alert_ref: {current_alert_ref_id}; updating stored alert_ref.")
+            set_alert_ref_id(redis_conn, posted_alert_ref)
+            current_alert_ref_id = posted_alert_ref
         src_image_filename = entity["filename"]["value"]
         src_image_bucket = entity["bucket"]["value"]
 
         # Submit tasks for both PersonVehicleDetection and FloodSegmentation
         # entities_to_explain = ['PersonVehicleDetection']
+        # entities_to_explain = ['FloodSegmentation']
         entities_to_explain = ['FloodSegmentation', 'PersonVehicleDetection']
 
         task_ids = []
@@ -272,6 +321,10 @@ def post_data():
                 'entity_type': entity_type,
                 'src_image_bucket': src_image_bucket,
                 'minio_filename': src_image_filename,
+                'uav_id': posted_uav_id,
+                'flight_number': posted_flight_number,
+                'bm_id': posted_bm_id,
+                'alert_ref': posted_alert_ref
             }
             update_job_status(redis_conn, task_id, job_status)
 
@@ -282,6 +335,10 @@ def post_data():
                 src_image_bucket,
                 src_image_filename,
                 task_id,
+                uav_id=posted_uav_id,
+                flight_number=posted_flight_number,
+                bm_id=posted_bm_id,
+                alert_ref=posted_alert_ref,
                 job_timeout='12h'  # Set an appropriate timeout
             )
             task_ids.append(task_id)
@@ -322,15 +379,26 @@ def post_data_old():
             app.logger.debug(f"Received outer entity type: {outer_entity_type} instead of Notification")
 
         if entity_type == "Alert":
+            uav_id = entity["uav_id"]["value"]
+            flight_number = entity["flight_number"]["value"]
             bm_id = entity["bm_id"]["value"]
-            set_bm_id(bm_id)
-            msg = f"Received Alert with bm_id: {bm_id}"
+            alert_ref = entity["alert_ref"]["value"]
+            set_alert_ref_id(redis_conn, alert_ref)
+            set_bm_id(redis_conn, bm_id)
+            set_uav_id(redis_conn, uav_id)
+            set_flight_number(redis_conn, flight_number)
+            msg = f"Received Alert with bm_id:  {bm_id} and alert_ref: {alert_ref} saved to redis"
             return jsonify({'message': msg}), 200
 
-        bm_id = get_bm_id()
+        bm_id = get_bm_id(redis_conn)
+        uav_id = get_uav_id(redis_conn)
+        flight_number = get_flight_number(redis_conn)
+        alert_ref = get_alert_ref_id(redis_conn)
+        app.logger.debug(f"Current alert_ref: {alert_ref}")
 
         app.logger.debug(f"Current bm_id: {bm_id}")
-
+        app.logger.debug(f"Current uav_id: {uav_id}")
+        app.logger.debug(f"Current flight_number: {flight_number}")
         explanator = get_explanator()
 
         if entity_type not in explanator.VALID_ENTITY_TYPES:
@@ -516,4 +584,4 @@ def subscribe_to_context_broker():
 # Run the Flask application
 if __name__ == '__main__':
     debug = os.environ.get('DEBUG', '').lower() in ('true', '1')
-    app.run(host='0.0.0.0', port=int(os.environ.get('PORT', 8080)), debug=debug)
+    app.run(host='0.0.0.0', port=int(os.environ.get('PORT', 8080)), debug=debug)

common_app_funcs.py

@@ -12,6 +12,30 @@ def get_bm_id(redis_conn):
     bm_id = redis_conn.get('current_bm_id')
     return bm_id.decode('utf-8') if bm_id else None
 
+def set_uav_id(redis_conn, uav_id):
+    redis_conn.set('current_uav_id', uav_id)
+
+def get_uav_id(redis_conn):
+    uav_id = redis_conn.get('current_uav_id')
+    return uav_id.decode('utf-8') if uav_id else None
+
+
+def set_flight_number(redis_conn, flight_number):
+    redis_conn.set('current_flight_number', flight_number)
+
+def get_flight_number(redis_conn):
+    flight_number = redis_conn.get('current_flight_number')
+    return flight_number.decode('utf-8') if flight_number else None
+
+
+
+def set_alert_ref_id(redis_conn, alert_ref):
+    redis_conn.set('current_alert_ref_id', alert_ref)
+
+def get_alert_ref_id(redis_conn):
+    alert_ref = redis_conn.get('current_alert_ref_id')
+    return alert_ref.decode('utf-8') if alert_ref else None
+
 def update_job_status(redis_conn, task_id, status_data):
     redis_conn.set(f"job_status:{task_id}", json.dumps(status_data))
 

configs/__init__.py

@@ -0,0 +1,11 @@
+# ------------------------------------------------------------------------------
+# Copyright (c) Microsoft
+# Licensed under the MIT License.
+# Written by Ke Sun (sunk@mail.ustc.edu.cn)
+# ------------------------------------------------------------------------------
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+from .default import _C as config
+from .default import update_config

configs/default.py

@@ -0,0 +1,105 @@
+# ------------------------------------------------------------------------------
+# Modified based on https://github.com/HRNet/HRNet-Semantic-Segmentation
+# ------------------------------------------------------------------------------
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+from yacs.config import CfgNode as CN
+
+
+_C = CN()
+
+_C.OUTPUT_DIR = ''
+_C.LOG_DIR = ''
+_C.GPUS = (0,)
+_C.WORKERS = 4
+_C.PRINT_FREQ = 20
+_C.AUTO_RESUME = False
+_C.PIN_MEMORY = True
+
+# Cudnn related params
+_C.CUDNN = CN()
+_C.CUDNN.BENCHMARK = True
+_C.CUDNN.DETERMINISTIC = False
+_C.CUDNN.ENABLED = True
+
+# common params for NETWORK
+_C.MODEL = CN()
+_C.MODEL.NAME = 'pidnet_s'
+_C.MODEL.PRETRAINED = 'pretrained_models/imagenet/PIDNet_S_ImageNet.pth.tar'
+_C.MODEL.ALIGN_CORNERS = True
+_C.MODEL.NUM_OUTPUTS = 2
+
+
+_C.LOSS = CN()
+_C.LOSS.USE_OHEM = True
+_C.LOSS.OHEMTHRES = 0.9
+_C.LOSS.OHEMKEEP = 100000
+_C.LOSS.CLASS_BALANCE = False
+_C.LOSS.BALANCE_WEIGHTS = [0.5, 0.5]
+_C.LOSS.SB_WEIGHTS = 0.5
+
+# DATASET related params
+_C.DATASET = CN()
+_C.DATASET.ROOT = 'data/'
+_C.DATASET.DATASET = 'cityscapes'
+_C.DATASET.NUM_CLASSES = 19
+_C.DATASET.TRAIN_SET = 'list/cityscapes/train.lst'
+_C.DATASET.EXTRA_TRAIN_SET = ''
+_C.DATASET.TEST_SET = 'list/cityscapes/val.lst'
+
+# training
+_C.FINETUNE = False
+_C.TRAIN = CN()
+_C.TRAIN.IMAGE_SIZE = [1024, 1024]  # width * height
+_C.TRAIN.BASE_SIZE = 2048
+_C.TRAIN.FLIP = True
+_C.TRAIN.MULTI_SCALE = True
+_C.TRAIN.SCALE_FACTOR = 16
+
+_C.TRAIN.LR = 0.01
+_C.TRAIN.EXTRA_LR = 0.001
+
+_C.TRAIN.OPTIMIZER = 'sgd'
+_C.TRAIN.MOMENTUM = 0.9
+_C.TRAIN.WD = 0.0001
+_C.TRAIN.NESTEROV = False
+_C.TRAIN.IGNORE_LABEL = -1
+
+_C.TRAIN.BEGIN_EPOCH = 0
+_C.TRAIN.END_EPOCH = 484
+_C.TRAIN.EXTRA_EPOCH = 0
+
+_C.TRAIN.RESUME = False
+
+_C.TRAIN.BATCH_SIZE_PER_GPU = 32
+_C.TRAIN.SHUFFLE = True
+
+# testing
+_C.TEST = CN()
+_C.TEST.IMAGE_SIZE = [2048, 1024]  # width * height
+_C.TEST.BASE_SIZE = 2048
+_C.TEST.BATCH_SIZE_PER_GPU = 32
+_C.TEST.MODEL_FILE = ''
+_C.TEST.FLIP_TEST = False
+_C.TEST.MULTI_SCALE = False
+
+_C.TEST.OUTPUT_INDEX = -1
+
+
+def update_config(cfg, args):
+    cfg.defrost()
+
+    cfg.merge_from_file(args.cfg)
+    cfg.merge_from_list(args.opts)
+
+    cfg.freeze()
+
+
+if __name__ == '__main__':
+    import sys
+    with open(sys.argv[1], 'w') as f:
+        print(_C, file=f)
+