From a4fa73deaafb1bbffc81ea99a171d12120c304ab Mon Sep 17 00:00:00 2001
From: Julius Unverfehrt <Julius.Unverfehrt@iqser.com>
Date: Wed, 23 Feb 2022 10:51:24 +0100
Subject: [PATCH] Pull request #14: optional debug progress bar added

Merge in RR/fb-detr from add-debug-progress-bar to master

Squashed commit of the following:

commit 3449be1b46f73a5e9ae3719ed2821a1b7faca9e4
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Wed Feb 23 10:26:47 2022 +0100

    refactoring; added VERBOSE flag to config

commit e50234e205dfd7a40aaf7981da85e28048d9efba
Merge: 89703ca f6c51be
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Wed Feb 23 09:45:33 2022 +0100

    Merge branch 'config_changes' into add-debug-progress-bar

commit f6c51beeaa952c18c80b7af6b7a46b9de8f521c3
Author: Matthias Bisping <matthias.bisping@iqser.com>
Date:   Wed Feb 23 09:44:00 2022 +0100

    added env var

commit 89703caa776f0fad55757ab22568e45949b2b310
Author: Julius Unverfehrt <Julius.Unverfehrt@iqser.com>
Date:   Wed Feb 23 08:28:52 2022 +0100

    optional debug progress bar added
---
 config.yaml          |  5 +++--
 fb_detr/predictor.py | 14 ++++++++++----
 src/serve.py         |  4 ----
 3 files changed, 13 insertions(+), 10 deletions(-)

diff --git a/config.yaml b/config.yaml
index 7569e6b..2a202b4 100644
--- a/config.yaml
+++ b/config.yaml
@@ -11,5 +11,6 @@ webserver:
   mode: $SERVER_MODE|production  # webserver mode: {development, production}
 
 service:
-  logging_level: DEBUG
-  batch_size: $BATCH_SIZE|2  # Number of images in memory simultaneously per service instance
+  logging_level: $LOGGING_LEVEL_ROOT|DEBUG  # Logging level for service logger
+  batch_size: $BATCH_SIZE|2  # Number of images in memory simultaneously
+  verbose: $VERBOSE|True  # Service prints document processing progress to stdout
diff --git a/fb_detr/predictor.py b/fb_detr/predictor.py
index a8f89eb..e083dbf 100644
--- a/fb_detr/predictor.py
+++ b/fb_detr/predictor.py
@@ -139,6 +139,15 @@ class Predictor:
         return predictions
 
     def predict_pdf(self, pdf: bytes):
+        def progress(generator):
+
+            page_count = get_page_count(pdf)
+            batch_count = int(page_count / CONFIG.service.batch_size)
+
+            yield from tqdm(
+                generator, total=batch_count, position=1, leave=True
+            ) if CONFIG.service.verbose else generator
+
         def predict_batch(batch_idx, batch):
             predictions = self.predict(batch)
             for p in predictions:
@@ -146,11 +155,8 @@ class Predictor:
 
             return predictions
 
-        page_count = get_page_count(pdf)
-        batch_count = int(page_count / CONFIG.service.batch_size)
-
         page_stream = stream_pages(pdf)
         page_batches = chunk_iterable(page_stream, CONFIG.service.batch_size)
-        predictions = list(chain(*starmap(predict_batch, tqdm(enumerate(page_batches), total=batch_count))))
+        predictions = list(chain(*starmap(predict_batch, progress(enumerate(page_batches)))))
 
         return predictions
diff --git a/src/serve.py b/src/serve.py
index c1c9deb..fa44ef9 100644
--- a/src/serve.py
+++ b/src/serve.py
@@ -1,16 +1,12 @@
 import argparse
-import json
 import logging
-from itertools import chain
 from typing import Callable
 
 from flask import Flask, request, jsonify
-from pdf2image import pdf2image
 from waitress import serve
 
 from fb_detr.config import CONFIG
 from fb_detr.utils.estimator import suppress_userwarnings, initialize_predictor
-from fb_detr.utils.stream import stream_pages, chunk_iterable
 
 
 def parse_args():