From f00a74a5648b122e2f946fc76c6fab0d59033f37 Mon Sep 17 00:00:00 2001
From: vladd-bit <vlad.a.dinu@gmail.com>
Date: Mon, 19 Feb 2024 11:00:19 +0000
Subject: [PATCH] Dockerfile update(s) + pep8 fix.

---
 Dockerfile                                    |  3 ++-
 Dockerfile_gpu                                |  9 ++++++--
 config.py                                     |  2 --
 .../nlp_processor/medcat_processor.py         | 23 ++++++++++---------
 4 files changed, 21 insertions(+), 16 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index 2f4c6c3..579dbdf 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -6,8 +6,9 @@ ENV CRYPTOGRAPHY_DONT_BUILD_RUST=1
 WORKDIR /cat
 COPY ./medcat_service/requirements.txt /cat
 RUN pip install --upgrade pip
+
 # Install requirements for the app
-RUN pip install -r requirements.txt
+RUN pip3 install --no-cache-dir -r requirements.txt
 
 # Get the spacy model
 ARG SPACY_MODELS="en_core_web_sm en_core_web_md en_core_web_lg"
diff --git a/Dockerfile_gpu b/Dockerfile_gpu
index 8847cbe..3bf2f9b 100644
--- a/Dockerfile_gpu
+++ b/Dockerfile_gpu
@@ -14,7 +14,6 @@ RUN echo "deb-src http://deb.debian.org/debian bookworm-backports main contrib n
 
 RUN apt-get update && apt-get upgrade -y && \
      apt-get install -y software-properties-common
-
 # NVIDIA Docker
 RUN export distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
 
@@ -26,13 +25,19 @@ RUN apt-get update && apt-get upgrade -y
 
 RUN apt-get install -y nvidia-container-toolkit nvidia-docker2
 
+# clean up apt
+RUN apt-get clean autoclean && apt-get autoremove --purge -y
+
 # Set the python path and preapre the base layer
 WORKDIR /cat
 COPY ./medcat_service/requirements.txt /cat
 RUN pip install --upgrade pip
 
+# clean up pip
+RUN pip3 cache purge
+
 # Install requirements for the app
-RUN pip install -r requirements.txt
+RUN pip install --no-cache-dir -r requirements.txt
 
 # Get the spacy model
 ARG SPACY_MODELS="en_core_web_sm en_core_web_md en_core_web_lg"
diff --git a/config.py b/config.py
index f6f8256..7c24459 100644
--- a/config.py
+++ b/config.py
@@ -16,5 +16,3 @@ def post_fork(server, worker):
         os.environ["CUDA_VISIBLE_DEVICES"] = str(cudaid)
     else:
         worker.log.info("APP_CUDA_DEVICE_COUNT device variables not set")
-
-
diff --git a/medcat_service/nlp_processor/medcat_processor.py b/medcat_service/nlp_processor/medcat_processor.py
index 4ba772f..a1c5340 100644
--- a/medcat_service/nlp_processor/medcat_processor.py
+++ b/medcat_service/nlp_processor/medcat_processor.py
@@ -19,6 +19,7 @@ class NlpProcessor:
     """
     This class defines an interface for NLP Processor
     """
+
     def __init__(self):
         app_log_level = os.getenv("APP_LOG_LEVEL", logging.INFO)
         medcat_log_level = os.getenv("LOG_LEVEL", logging.INFO)
@@ -114,10 +115,10 @@ def process_content(self, content):
         if "text" not in content:
             error_msg = "'text' field missing in the payload content."
             nlp_result = {
-                          "success": False,
-                          "errors": [error_msg],
-                          "timestamp": NlpProcessor._get_timestamp(),
-                         }
+                "success": False,
+                "errors": [error_msg],
+                "timestamp": NlpProcessor._get_timestamp(),
+            }
 
             return nlp_result
 
@@ -138,12 +139,12 @@ def process_content(self, content):
         entities = self.process_entities(entities)
 
         nlp_result = {
-                      "text": str(text),
-                      "annotations": entities,
-                      "success": True,
-                      "timestamp": NlpProcessor._get_timestamp(),
-                      "elapsed_time":  elapsed_time
-                      }
+            "text": str(text),
+            "annotations": entities,
+            "success": True,
+            "timestamp": NlpProcessor._get_timestamp(),
+            "elapsed_time":  elapsed_time
+        }
 
         # append the footer
         if "footer" in content:
@@ -187,7 +188,7 @@ def process_content_bulk(self, content):
                 ann_res = self.cat.deid_text()
             else:
                 ann_res = self.cat.multiprocessing_batch_docs_size(
-                    MedCatProcessor._generate_input_doc(content, invalid_doc_ids), batch_size = batch_size, nproc=nproc)
+                    MedCatProcessor._generate_input_doc(content, invalid_doc_ids), batch_size=batch_size, nproc=nproc)
 
         except Exception as e:
             self.log.error(repr(e))