Scraper for .nfo exports from kodi/plex

scrapers/kodi.py

@@ -0,0 +1,113 @@
+import os
+import sys
+import json
+import sqlite3
+import mimetypes
+import base64
+import xml.etree.ElementTree as ET
+"""  
+This script parses kodi nfo files for metadata. The .nfo file must be in the same directory as the video file and must be named exactly alike.
+"""
+debug = False
+
+
+# If you want to ingest image files from the .nfo the path to these files may need to be rewritten. Especially when using a docker container.
+rewriteBasePath = False
+# Example: Z:\Videos\Studio_XXX\example_cover.jpg -> /data/Studio_XXX/example_cover.jpg
+basePathBefore = 'Z:\Videos'
+basePathAfter = "/data"
+
+def query_xml(path, title):
+    tree=ET.parse(path)
+    # print(tree.find("title").text, file=sys.stderr)
+    if title == tree.find("title").text:
+        debug("Exact match found for " + title)
+    else:
+        debug("No exact match found for " + title + ". Matching with " + tree.find("title").text + "!")
+
+    # Extract matadata from xml
+    res={"title":title}
+    if tree.find("title") != None:
+        res["title"] = tree.find("title").text
+    if tree.find("plot") != None:
+        res["details"] = tree.find("plot").text
+    if tree.find("releasedate") != None:
+        res["date"] = tree.find("releasedate").text
+    if tree.find("tag") != None:
+        res["tags"]=[{"name":x.text} for x in tree.findall("tag")]
+    if tree.find("genre") != None:
+        if "tags" in res:
+            res["tags"] += [{"name":x.text} for x in tree.findall("genre")]
+        else:
+            res["tags"] = [{"name":x.text} for x in tree.findall("genre")]
+    if tree.find("actor") != None:
+        res["performers"] = []
+        for actor in tree.findall("actor"):
+            if actor.find("type") != None:
+                if actor.find("type").text == "Actor":
+                    res["performers"].append({"name": actor.find("name").text})
+            else if actor.find("name") != None:
+                res["performers"].append({"name": actor.find("name").text})
+            else:
+                res["performers"].append({"name": actor.text})
+    if tree.find("studio") != None:
+        res["studio"] = {"name":tree.find("studio").text}
+
+    if tree.find("art") != None:
+        if tree.find("art").find("poster") != None:
+            posterElem = tree.find("art").find("poster")
+            if posterElem.text != None:
+                if not rewriteBasePath and os.path.isfile(posterElem.text):
+                    res["image"] = make_image_data_url(posterElem.text)
+                elif rewriteBasePath:
+                    rewrittenPath = posterElem.text.replace(basePathBefore, basePathAfter).replace("\\", "/")
+                    if os.path.isfile(rewrittenPath):
+                        res["image"] = make_image_data_url(rewrittenPath)
+                    else:
+                        debug("Can't find image: " + posterElem.text.replace(basePathBefore, basePathAfter) + ". Is the base path correct?")
+                else:
+                    debug("Can't find image: " + posterElem.text + ". Are you using a docker container? Maybe you need to change the base path in the script file.")
+
+    return res
+
+def debug(s):
+    if debug: print(s, file=sys.stderr)
+
+# Would be nicer with Stash API instead of direct SQlite access
+def get_file_path(scene_id):
+    db_file = "../stash-go.sqlite"
+
+    con = sqlite3.connect(db_file)
+    cur = con.cursor()
+    for row in cur.execute("SELECT * FROM scenes where id = " + str(scene_id) + ";"):
+        #debug_print(row)
+        filepath = row[1]
+    con.close()
+    return filepath
+
+def make_image_data_url(image_path):
+    # type: (str,) -> str
+    mime, _ = mimetypes.guess_type(image_path)
+    with open(image_path, 'rb') as img:
+        encoded = base64.b64encode(img.read()).decode()
+    return 'data:{0};base64,{1}'.format(mime, encoded)
+
+if sys.argv[1] == "query":
+    fragment = json.loads(sys.stdin.read())
+    res = {"title": fragment["title"]}
+    # Assume that .nfo is named exactly like the video file and is at the same location
+    # WORKAROUND: Read file name from db until filename is given in the fragment
+    videoFilePath = get_file_path(fragment["id"])
+
+    # Reconstruct file name for .nfo
+    temp = videoFilePath.split(".")
+    temp[-1] = "nfo"
+    nfoFilePath = ".".join(temp)
+
+    if os.path.isfile(nfoFilePath):
+        res = query_xml(nfoFilePath, fragment["title"])
+    else:
+        debug("No file found at" + nfoFilePath)
+
+    print(json.dumps(res))
+    exit(0)

scrapers/kodi.yml

@@ -0,0 +1,10 @@
+name: "Kodi XML"
+sceneByFragment:
+    action: script
+    script:
+      - python
+      # use python3 instead if needed
+      - kodi.py
+      - query
+
+# Last Updated August 15, 2021