c668228fe8
chore: specify too broad exception types
2024-02-04 01:31:24 +01:00
8af73e622f
chore: remove remaining todos after analysis
...
refiner is added when set
restoring multiline prompts has been resolved by using separate parameters "raw_prompt" and "raw_negative_prompt"
2024-02-04 00:44:26 +01:00
1419231e74
feat: prevent config reset by renaming metadata_scheme to match config options
2024-02-03 22:41:26 +01:00
63403d614e
feat: add sampler mapping
2024-02-02 23:44:47 +01:00
ed4a958da8
fix: add workaround for multiline prompts
2024-02-02 22:04:28 +01:00
349556bfa6
fix: remove sorting for A1111, change performance key position
...
fixes https://github.com/lllyasviel/Fooocus/pull/1940#issuecomment-1924444633
2024-02-02 20:58:16 +01:00
1c3431e6a7
wip: add TODO for ultiline prompt style resolution
2024-02-02 02:02:41 +01:00
9aa82aa80a
fix: add missing refiner in A1111 parse_json
2024-02-02 01:57:33 +01:00
f745d40687
refactor: merge metadata.py into meta_parser.py
2024-02-02 01:55:32 +01:00
b438f7b43f
chore: check and remove ToDo's
2024-02-02 01:45:08 +01:00
f7e24bdb1f
feat: sort metadata attributes before writing to image
2024-02-02 01:26:20 +01:00
e55870124b
refactor: add step before parsing to set data in parser
...
- add constructor for MetadataSchema class
- remove showable and copyable from log output
- add functional hash cache (model hashing takes about 5 seconds, only required once per model, using hash lazy loading)
2024-02-02 01:25:47 +01:00
796cf3c78d
fix: align handling for scheme Fooocus by removing lcm lora from json parsing
2024-02-02 01:20:28 +01:00
f4afc4af10
feat: add method get_scheme to Metadata
2024-02-02 01:19:21 +01:00
ea6839be83
chore: code cleanup
2024-02-02 01:14:40 +01:00
bc9b625221
feat: add hash cache to metadata
2024-02-02 01:11:50 +01:00
23ba05015a
chore: use str as return type in calculate_sha256
2024-02-02 01:10:26 +01:00
6b9c0bd448
refactor: code cleanup
2024-01-31 01:35:51 +01:00
9bdb65ec5d
feat: add metadata handling for all non-img2img parameters
2024-01-31 01:18:09 +01:00
7772eb7965
refactor: add flag for refiner_swap_method
2024-01-31 01:17:10 +01:00
e93a345ccf
feat: code cleanup, do not queue metadata preview on image upload
2024-01-30 00:20:46 +01:00
5811234d23
feat: allow model / LoRA loading from subfolders
2024-01-30 00:01:29 +01:00
6939f7904a
fix: correctly extract prompt style when only prompt expansion is selected
2024-01-29 22:27:33 +01:00
dcc4874455
feat: override existing LoRA values when importing, would cause images to differ
2024-01-29 21:57:02 +01:00
78d1ad3962
feat: add lora handling to A1111 scheme
2024-01-29 21:56:10 +01:00
89c8e3a812
feat: make sha256 with length 10 default
2024-01-29 21:54:39 +01:00
e541097451
wip: code cleanup, update todos
2024-01-29 21:54:22 +01:00
2656356206
fix: use correct order to read LoRA in meta parser
2024-01-29 18:17:51 +01:00
e388f6fff6
wip: code cleanup, update todos
2024-01-29 16:38:39 +01:00
33d644f4a5
feat: add prefix "Fooocus" to version
2024-01-29 16:29:40 +01:00
7fefe3a3c2
feat: add created_by again
2024-01-29 16:28:47 +01:00
c80011b1d1
fix: use correct LoRA mapping, add fallback for backwards compatibility
2024-01-29 15:45:55 +01:00
20e53028a4
refactor: use central flag for ControlNet image count
2024-01-29 14:27:51 +01:00
c3ab9f1f30
refactor: use central flag for LoRA count
2024-01-29 14:26:56 +01:00
f94b96f6eb
wip: add prompt style extraction for A1111 scheme
2024-01-29 01:52:24 +01:00
5e84a45e22
feat: add A1111 prompt style detection
...
only detects one style as Fooocus doesn't wrap {prompt} with the whole style, but has a separate prompt string for each style
2024-01-28 23:52:06 +01:00
236278948b
chore: code cleanup
2024-01-28 23:50:56 +01:00
5dcb2bc573
fix: correctly map resolution, use empty styles for A1111
2024-01-28 20:42:58 +01:00
cbc63ebba3
feat: add enums for Performance, Steps and StepsUOV
...
also move MetadataSchema enum to prevent circular dependency
2024-01-28 20:01:33 +01:00
e19596c2df
feat: map basic information for scheme A1111
2024-01-28 18:04:40 +01:00
f3010313fc
wip: add metadata mapping, reading and writing
...
applying data after reading currently not functional for A1111
2024-01-28 05:35:44 +01:00
051faf78b8
fix: use correct default value in metadata check for created_by
2024-01-25 23:49:25 +01:00
20b79788a0
feat: add resolved prompts to metadata
2024-01-25 23:48:47 +01:00
510b587e84
fix: use pnginfo "parameters" insteadf of "Comments"
...
see https://github.com/RupertAvery/DiffusionToolkit/issues/202 and cf2772fab0/modules/processing.py (L939)
2024-01-25 21:08:19 +01:00
ba5d0b6eaf
refactor: rename metadata_schema to metadata_scheme
2024-01-15 23:59:57 +01:00
80ad0d070e
fix: do not generate hashes when arg --disable-metadata is used
2024-01-15 23:58:34 +01:00
cd65f21d98
fix: use stting isntead of quote wrap for A1111 created_by
2024-01-15 23:14:49 +01:00
7b9deb17ee
feat: add config metadata_created_by
2024-01-15 23:06:20 +01:00
6662381931
fix: code cleanup and resolved prompt fixes
2024-01-15 23:00:59 +01:00
1a52367f3f
feat: use resolved prompts with included expansion and styles for a1111 metadata
2024-01-15 22:17:51 +01:00
Manuel Schmid
Manuel Schmid