Use source hashes for cover matching

comicapi/genericmetadata.py

@@ -136,6 +136,14 @@ class MetadataOrigin(NamedTuple):
         return self.name
 
 
+class ImageHash(NamedTuple):
+    Hash: int
+    Kind: str  # ahash, phash
+
+    def __str__(self) -> str:
+        return str(self.Hash) + ": " + self.Kind
+
+
 @dataclasses.dataclass
 class GenericMetadata:
     writer_synonyms = ("writer", "plotter", "scripter", "script")
@@ -202,7 +210,7 @@ class GenericMetadata:
     last_mark: str | None = None
 
     # urls to cover image, not generally part of the metadata
-    _cover_image: str | None = None
+    _cover_image: str | ImageHash | None = None
     _alternate_images: list[str] = dataclasses.field(default_factory=list)
 
     def __post_init__(self) -> None:

comictaggerlib/cli.py

@@ -639,7 +639,7 @@ class CLI:
                             month=ct_md.month,
                             year=ct_md.year,
                             publisher=None,
-                            image_url=ct_md._cover_image or "",
+                            image_url=str(ct_md._cover_image) or "",
                             alt_image_urls=[],
                             description=ct_md.description or "",
                         )

comictaggerlib/issueidentifier.py

@@ -25,7 +25,7 @@ from typing_extensions import NotRequired, TypedDict
 
 from comicapi import utils
 from comicapi.comicarchive import ComicArchive
-from comicapi.genericmetadata import ComicSeries, GenericMetadata
+from comicapi.genericmetadata import ComicSeries, GenericMetadata, ImageHash
 from comicapi.issuestring import IssueString
 from comictaggerlib.ctsettings import ct_ns
 from comictaggerlib.imagefetcher import ImageFetcher, ImageFetcherException
@@ -132,13 +132,13 @@ class IssueIdentifier:
     def set_cover_url_callback(self, cb_func: Callable[[bytes], None]) -> None:
         self.cover_url_callback = cb_func
 
-    def calculate_hash(self, image_data: bytes) -> int:
+    def calculate_hash(self, image_data: bytes = b"", image: Image = None) -> int:
         if self.image_hasher == 3:
-            return ImageHasher(data=image_data).p_hash()
+            return ImageHasher(data=image_data, image=image).p_hash()
         if self.image_hasher == 2:
-            return -1  # ImageHasher(data=image_data).average_hash2()
+            return -1  # ImageHasher(data=image_data, image=image).average_hash2()
 
-        return ImageHasher(data=image_data).average_hash()
+        return ImageHasher(data=image_data, image=image).average_hash()
 
     def log_msg(self, msg: Any) -> None:
         msg = str(msg)
@@ -306,7 +306,7 @@ class IssueIdentifier:
 
     def _get_issue_cover_match_score(
         self,
-        primary_img_url: str,
+        primary_img_url: str | ImageHash,
         alt_urls: list[str],
         local_hashes: list[tuple[str, int]],
         use_alt_urls: bool = False,
@@ -320,12 +320,15 @@ class IssueIdentifier:
 
         self._user_canceled()
 
-        urls = [primary_img_url]
-        if use_alt_urls:
-            urls.extend(alt_urls)
-            self.log_msg(f"[{len(alt_urls)} alt. covers]")
+        if isinstance(primary_img_url, ImageHash):
+            remote_hashes = [("0", primary_img_url.Hash)]
+        else:
+            urls = [primary_img_url]
+            if use_alt_urls:
+                urls.extend(alt_urls)
+                self.log_msg(f"[{len(alt_urls)} alt. covers]")
 
-        remote_hashes = self._get_remote_hashes(urls)
+            remote_hashes = self._get_remote_hashes(urls)
 
         score_list = []
         done = False
@@ -490,7 +493,7 @@ class IssueIdentifier:
     def _calculate_hashes(self, images: list[tuple[str, Image.Image]]) -> list[tuple[str, int]]:
         hashes = []
         for name, image in images:
-            hashes.append((name, ImageHasher(image=image).average_hash()))
+            hashes.append((name, self.calculate_hash(image=image)))
         return hashes
 
     def _match_covers(
@@ -536,7 +539,7 @@ class IssueIdentifier:
                 month=issue.month,
                 year=issue.year,
                 publisher=None,
-                image_url=image_url,
+                image_url=str(image_url),
                 alt_image_urls=alt_urls,
                 description=issue.description or "",
             )
@@ -620,6 +623,16 @@ class IssueIdentifier:
         extra_images: list[tuple[str, Image.Image]],
         issues: list[tuple[ComicSeries, GenericMetadata]],
     ) -> list[IssueResult]:
+        # Set hashing kind, will presume all hashes are of the same kind
+        for series, issue in issues:
+            if isinstance(issue._cover_image, ImageHash):
+                if issue._cover_image.Kind == "phash":
+                    self.image_hasher = 3
+                    break
+                elif issue._cover_image == "ahash":
+                    self.image_hasher = 1  # Set to 1 on init but might as well be sure
+                    break
+
         cover_matching_1 = self._match_covers(terms, images, issues, use_alternates=False)
 
         if len(cover_matching_1) == 0:

comictaggerlib/issueselectionwindow.py

@@ -221,7 +221,10 @@ class IssueSelectionWindow(QtWidgets.QDialog):
         QtWidgets.QApplication.restoreOverrideCursor()
 
         self.issue_number = issue.issue or ""
-        self.coverWidget.set_issue_details(self.issue_id, [issue._cover_image or "", *issue._alternate_images])
+        cover_image = ""
+        if isinstance(issue._cover_image, str):
+            cover_image = issue._cover_image
+        self.coverWidget.set_issue_details(self.issue_id, [cover_image, *issue._alternate_images])
         if issue.description is None:
             self.set_description(self.teDescription, "")
         else: