Reuse code for metadata

NCISlideUtil.py

@@ -72,7 +72,7 @@ def openslidedata(metadata):
     slide = openslide.OpenSlide(metadata['location'])
     slideData = slide.properties
     metadata['mpp-x'] = slideData.get(openslide.PROPERTY_NAME_MPP_X, None)
-    metadata['mpp-x'] = slideData.get(openslide.PROPERTY_NAME_MPP_Y, None)
+    metadata['mpp-y'] = slideData.get(openslide.PROPERTY_NAME_MPP_Y, None)
     metadata['mpp'] = metadata['mpp-x'] or metadata['mpp-x'] or None
     metadata['height'] = slideData.get(
         openslide.PROPERTY_NAME_BOUNDS_HEIGHT, None)

OmniLoad.py

@@ -8,6 +8,7 @@
 import json # for json in and out
 import requests # for api and pathdb in and out
 import hashlib
+import dev_utils
 
 # for large csv fields, especially segmentations
 csv.field_size_limit(sys.maxsize)
@@ -51,21 +52,10 @@ def file_md5(fileName):
 def openslidedata(manifest):
     for img in manifest:
         img['location'] = img.get("path", "") or img.get("location", "") or img.get("filename", "") or img.get("file", "")
-        slide = openslide.OpenSlide(img['location'])
-        slideData = slide.properties
-        img['mpp-x'] = slideData.get(openslide.PROPERTY_NAME_MPP_X, None)
-        img['mpp-y'] = slideData.get(openslide.PROPERTY_NAME_MPP_Y, None)
-        img['mpp'] = img['mpp-x'] or img['mpp-y']
-        img['height'] = slideData.get(openslide.PROPERTY_NAME_BOUNDS_HEIGHT, None) or slideData.get(
-            "openslide.level[0].height", None)
-        img['width'] = slideData.get(openslide.PROPERTY_NAME_BOUNDS_WIDTH, None) or slideData.get(
-            "openslide.level[0].width", None)
-        img['vendor'] = slideData.get(openslide.PROPERTY_NAME_VENDOR, None)
-        img['level_count'] = int(slideData.get('level_count', 1))
-        img['objective'] = float(slideData.get(openslide.PROPERTY_NAME_OBJECTIVE_POWER, 0) or
-                                      slideData.get("aperio.AppMag", -1.0))
-        img['md5sum'] = file_md5(img['location'])
-        img['comment'] = slideData.get(openslide.PROPERTY_NAME_COMMENT, None)
+        metadata = dev_utils.getMetadata(img['location'], False, True)
+        for k, v in metadata.items():
+          if k not in img:
+            img[k] = v
         # required values which are often unused
         img['study'] = img.get('study', "")
         img['specimen'] = img.get('specimen', "")

SlideServer.py

@@ -6,8 +6,7 @@
 import string
 import sys
 import pyvips
-from os import listdir
-from os.path import isfile, join
+import os
 from spritemaker import createSpritesheet
 from PIL import Image
 import urllib
@@ -203,7 +202,7 @@ def testRoute():
 @app.route("/data/one/<filepath>", methods=['GET'])
 def singleSlide(filepath):
     extended = request.args.get('extended')
-    res = dev_utils.getMetadata(filepath, app.config['UPLOAD_FOLDER'], extended)
+    res = dev_utils.getMetadata(os.path.join(app.config['UPLOAD_FOLDER'], filepath), extended, False)
     if (hasattr(res, 'error')):
         return flask.Response(json.dumps(res), status=500)
     else:
@@ -223,8 +222,10 @@ def singleThumb(filepath):
 
 @app.route("/data/many/<filepathlist>", methods=['GET'])
 def multiSlide(filepathlist):
-    request.args.get('extended')
-    res = dev_utils.getMetadataList(json.loads(filepathlist), app.config['UPLOAD_FOLDER'], extended)
+    extended = request.args.get('extended')
+    filenames = json.loads(filepathlist)
+    paths = [os.path.join(app.config['UPLOAD_FOLDER'], filename) for filename in filenames]
+    res = dev_utils.getMetadataList(paths, extended, False)
     if (hasattr(res, 'error')):
         return flask.Response(json.dumps(res), status=500)
     else:

SlideUtil.py

@@ -5,7 +5,7 @@
 
 import openslide
 
-from dev_utils import file_md5
+from dev_utils import getMetadata
 from dev_utils import postslide
 from dev_utils import post_url
 
@@ -34,22 +34,14 @@ def gen_thumbnail(filename, slide, size, imgtype="png"):
 
 
 def openslidedata(metadata):
-    slide = openslide.OpenSlide(metadata['location'])
-    slideData = slide.properties
-    metadata['mpp-x'] = slideData.get(openslide.PROPERTY_NAME_MPP_X, None)
-    metadata['mpp-x'] = slideData.get(openslide.PROPERTY_NAME_MPP_Y, None)
-    metadata['mpp'] = metadata['mpp-x'] or metadata['mpp-x'] or None
-    # metadata['height'] = slideData.get("openslide.level[0].height", None)
-    # metadata['width'] = slideData.get("openslide.level[0].width", None)
-    metadata['height'] = slideData.get(openslide.PROPERTY_NAME_BOUNDS_HEIGHT, None)
-    metadata['width'] = slideData.get(openslide.PROPERTY_NAME_BOUNDS_WIDTH, None)
-    metadata['vendor'] = slideData.get(openslide.PROPERTY_NAME_VENDOR, None)
-    metadata['level_count'] = int(slideData.get('level_count', 1))
-    metadata['objective'] = float(slideData.get("aperio.AppMag", 0.0))
-    metadata['md5sum'] = file_md5(metadata['location'])
+    metadata_retrieved = getMetadata(metadata['location'], False, True)
+    for k, v in metadata_retrieved.items():
+        if k not in metadata:
+            metadata[k] = v
     metadata['timestamp'] = time.time()
     thumbnail_size = config.get('thumbnail_size', None)
     if thumbnail_size:
+        slide = openslide.OpenSlide(metadata['location'])
         gen_thumbnail(metadata['location'], slide, thumbnail_size)
     return metadata
 

dev_utils.py

@@ -10,18 +10,21 @@
 
 
 # given a path, get metadata
-def getMetadata(filename, upload_folder, extended):
+def getMetadata(filepath, extended, raise_exception):
     # TODO consider restricting filepath
     metadata = {}
-    filepath = os.path.join(upload_folder, filename)
     if not os.path.isfile(filepath):
+        if raise_exception:
+            raise ValueError("No such file")
         msg = {"error": "No such file"}
         print(msg)
         return msg
     metadata['location'] = filepath
     try:
         slide = openslide.OpenSlide(filepath)
     except BaseException as e:
+        if raise_exception:
+            raise e
         msg = {"type": "Openslide", "error": str(e)}
         print(msg)
         return msg
@@ -60,10 +63,10 @@ def postslide(img, url, token=''):
 
 
 # given a list of path, get metadata for each
-def getMetadataList(filenames, upload_folder, extended):
+def getMetadataList(filenames, extended, raise_exception):
     allData = []
     for filename in filenames:
-        allData.append(getMetadata(filename, upload_folder, extended))
+        allData.append(getMetadata(filename, extended, raise_exception))
     return allData