Merge pull request #230 from tattle-made/development

merge dev to main
tattle-made · Mar 21, 2024 · ed75278 · ed75278
2 parents 34f3ed0 + 81a99ed
commit ed75278
Show file tree

Hide file tree

Showing 14 changed files with 495 additions and 38 deletions.
diff --git a/docker-compose.yml b/docker-compose.yml
@@ -3,7 +3,7 @@ version: "3.5"
 services:
   store:
     container_name: es
-    image: docker.elastic.co/elasticsearch/elasticsearch:8.12.0
+    image: docker.elastic.co/elasticsearch/elasticsearch@sha256:ec72548cf833e58578d8ff40df44346a49480b2c88a4e73a91e1b85ec7ef0d44 # docker.elastic.co/elasticsearch/elasticsearch:8.12.0
     volumes:
       - ./.docker/es/data:/var/lib/elasticsearch/data
     ports:
@@ -27,7 +27,7 @@ services:
       - IPC_LOCK
 
   queue:
-    image: rabbitmq:3.12.12-management
+    image: rabbitmq@sha256:e2505f78d58dca8c372cde3930e4d6bee5e02ac9001ce85ece5a11df606c1fa3 # rabbitmq:3.12.12-management
     container_name: rabbitmq
     hostname: rabbit
     volumes:
@@ -52,6 +52,7 @@ services:
         - "GID=${GID:-1000}"
     volumes:
       - ./src:/usr/app
+      - venv:/usr/app/venv/
     env_file: ./src/development.env
     ports:
       - 7000:7000
@@ -69,6 +70,9 @@ services:
       context: ./src
       dockerfile: Dockerfile
       target: debug
+      args:
+        - "UID=${UID:-1000}"
+        - "GID=${GID:-1000}"
     env_file: ./src/development.env
     command: tail -f /dev/null
     depends_on:
@@ -83,6 +87,9 @@ services:
       context: ./src
       dockerfile: Dockerfile
       target: debug
+      args:
+        - "UID=${UID:-1000}"
+        - "GID=${GID:-1000}"
     env_file: ./src/development.env
     command: tail -f /dev/null
     depends_on:
@@ -91,6 +98,9 @@ services:
       queue:
         condition: service_started
 
+volumes:
+    venv: {}
+
 #networks:
 #  default:
 #    external: true

diff --git a/src/.env-template b/src/.env-template
@@ -13,4 +13,8 @@ ES_IMG_INDEX=imgsearch
 ES_TXT_INDEX=txtsearch
 ES_VID_INDEX=vidsearch
 WSGI_HOST=XXXXX
-WSGI_DEBUG=XXXXX
+WSGI_DEBUG=XXXXX
+PG_HOST=postgres
+PG_DB=postgres
+PG_USER=tattle
+PG_PASS=tattle_pw
diff --git a/src/Dockerfile b/src/Dockerfile
@@ -16,9 +16,9 @@ RUN apt-get update \
     && rm -rf /var/lib/apt/lists/* 
 
 # Set python user
-RUN groupadd -g $GID python \
-    && useradd --create-home -r -u $UID -g python python \
-    && mkdir /usr/app \
+RUN groupadd -f -g $GID python \
+    && id -u python >/dev/null 2>&1 || useradd --create-home -r -u $UID -g python python \
+    && mkdir -p /usr/app \
     && chown -R python:python /usr/app
 WORKDIR /usr/app
 
@@ -47,9 +47,27 @@ EXPOSE 7000
 
 #### DEBUG IMAGE ####
 FROM base AS debug
+ARG UID
+ARG GID
+
 RUN apt-get update && apt-get install -y --no-install-recommends vim zsh jq
+
+# Set python user
+RUN groupadd -f -g $GID python \
+    && id -u python >/dev/null 2>&1 || useradd --create-home -r -u $UID -g python python \
+    && mkdir -p /usr/app \
+    && chown -R python:python /usr/app
+WORKDIR /usr/app
+
+# Set venv path
+ENV PATH="/usr/app/venv/bin:$PATH"
+
+# install python packages
 RUN pip install --no-cache-dir debugpy nose2
+
+# Set unprivileged user with group membership
 USER python:python
+
 CMD python -m debugpy --listen localhost:5678 --wait-for-client -m flask run --debug -h localhost -p 5000
 
 #### PROD IMAGE ####

diff --git a/src/core/models/media_factory.py b/src/core/models/media_factory.py
@@ -9,6 +9,7 @@
 import logging
 import os
 import tempfile
+import boto3
 
 log = logging.getLogger(__name__)
 
@@ -68,21 +69,53 @@ def make_from_file_on_disk(image_path):
     def make_from_file_in_memory(image_data: FileStorage):
         pass
 
-
 class VideoFactory:
+    aws_access_key_id = os.getenv('AWS_ACCESS_KEY_ID')
+    aws_secret_access_key = os.getenv('AWS_SECRET_ACCESS_KEY')
+    aws_region = os.getenv('AWS_REGION')
+    aws_bucket = os.getenv('AWS_BUCKET')
+    session = boto3.Session(
+        aws_access_key_id=aws_access_key_id,
+        aws_secret_access_key=aws_secret_access_key,
+        region_name=aws_region
+    )
+    s3 = session.client('s3')
     @staticmethod
-    def make_from_url(video_url):
-        temp_dir = tempfile.gettempdir()
-        temp_url = video_url.split("?")[0]
-        file_name = temp_url.split("/")[-1] + ".mp4"
-        file_path = temp_dir + os.sep + file_name
+    def download_file_from_s3(bucket_name, file_key, local_file_path):
         try:
-            print("Downloading video from url")
-            wget.download(video_url, out=file_path)
-            print("video downloaded")
+            VideoFactory.s3.download_file(bucket_name, file_key, local_file_path)
+            print(f"File {file_key} downloaded successfully as {local_file_path}")
         except Exception as e:
-            log.exception("Error downloading video:", e)
-            raise Exception("Error Downloading Video")
+            print(f"Error downloading file {file_key}: {e}")
+
+    @staticmethod
+    def make_from_url(video_url):
+        temp_dir = tempfile.gettempdir()
+
+        if video_url.startswith("http"):
+            temp_url = video_url.split("?")[0]
+            file_name = temp_url.split("/")[-1] + ".mp4"
+            file_path = os.path.join(temp_dir, file_name)
+            try:
+                print("Downloading video from URL")
+                wget.download(video_url, out=file_path)
+                print("Video downloaded")
+            except Exception as e:
+                print("Error downloading video:", e)
+                raise Exception("Error Downloading Video")
+        else:
+            bucket_name = VideoFactory.aws_bucket
+            file_key = video_url
+            file_name = file_key.split("/")[-1]
+            file_path = os.path.join(temp_dir, file_name)
+            try:
+                print("Downloading video from S3")
+                VideoFactory.download_file_from_s3(bucket_name, file_key, file_path)
+                print("Video downloaded")
+            except Exception as e:
+                print("Error downloading video from S3:", e)
+                raise Exception("Error Downloading Video")
+
         return {"path": file_path}
 
     @staticmethod

diff --git a/src/core/operators/audio_vec_embedding.py b/src/core/operators/audio_vec_embedding.py
@@ -2,13 +2,16 @@ def initialize(param):
     global model
     global librosa
     global np
+    global contextmanager
+    global os
 
     import numpy as np
     import librosa
 
     # from panns_inference import AudioTagging
     from core.operators.audio_cnn_model.inference import AudioTagging
-    # import os
+    from contextlib import contextmanager
+    import os
 
     # load the default model into cpu.
     model = AudioTagging(checkpoint_path=None, device="cpu")
@@ -24,16 +27,25 @@ def normalize(v):
 
 def run(audio_file):
     audio = audio_file["path"]
-    a, _ = librosa.load(audio, sr=44100)
-    query_audio = a[None, :]
-    _, emb = model.inference(query_audio)
-    normalized_v = normalize(emb[0])
-    return normalized_v
+
+    @contextmanager
+    def audio_load(fname):
+        a, _ = librosa.load(fname, sr=44100)
+        try:
+            yield a
+        finally:
+            os.remove(fname)
+
+    with audio_load(audio) as audio_var:
+        query_audio = audio_var[None, :]
+        _, emb = model.inference(query_audio)
+        normalized_v = normalize(emb[0])
+        return normalized_v
 
 
 # if __name__ == "__main__":
 #     initialize(param={})
-#     audio_file_path = {"path": r'sample_data/audio.wav'}
+#     audio_file_path = {"path": r"core/operators/sample_data/audio.wav"}
 #     audio_emb = run(audio_file_path)
 #     audio_emb_list = audio_emb.tolist()
 #     print(len(audio_emb_list))
diff --git a/src/core/operators/audio_vec_embedding_requirements.txt b/src/core/operators/audio_vec_embedding_requirements.txt
@@ -1049,7 +1049,7 @@ uc-micro-py==1.0.3 \
     --hash=sha256:d321b92cff673ec58027c04015fcaa8bb1e005478643ff4a500882eaab88c48a \
     --hash=sha256:db1dffff340817673d7b466ec86114a9dc0e9d4d9b5ba229d9d60e5c12600cd5
     # via linkify-it-py
-urllib3==2.2.0 \
-    --hash=sha256:051d961ad0c62a94e50ecf1af379c3aba230c66c710493493560c0c223c49f20 \
-    --hash=sha256:ce3711610ddce217e6d113a2732fafad960a03fd0318c91faa79481e35c11224
+urllib3==2.2.1 \
+    --hash=sha256:450b20ec296a467077128bff42b73080516e71b56ff59a60a02bef2232c4fa9d \
+    --hash=sha256:d0570876c61ab9e520d776c38acbbb5b05a776d3f9ff98a5c8fd5162a444cf19
     # via requests
diff --git a/src/core/operators/vid_vec_rep_resnet_requirements.txt b/src/core/operators/vid_vec_rep_resnet_requirements.txt
@@ -531,7 +531,7 @@ uc-micro-py==1.0.2 \
     --hash=sha256:30ae2ac9c49f39ac6dce743bd187fcd2b574b16ca095fa74cd9396795c954c54 \
     --hash=sha256:8c9110c309db9d9e87302e2f4ad2c3152770930d88ab385cd544e7a7e75f3de0
     # via linkify-it-py
-urllib3==2.2.0 \
-    --hash=sha256:051d961ad0c62a94e50ecf1af379c3aba230c66c710493493560c0c223c49f20 \
-    --hash=sha256:ce3711610ddce217e6d113a2732fafad960a03fd0318c91faa79481e35c11224
+urllib3==2.2.1 \
+    --hash=sha256:450b20ec296a467077128bff42b73080516e71b56ff59a60a02bef2232c4fa9d \
+    --hash=sha256:d0570876c61ab9e520d776c38acbbb5b05a776d3f9ff98a5c8fd5162a444cf19
     # via requests