feature: interactive prompt

- add quiet flag
- add mask-modify-original flag

README.md

@@ -121,6 +121,7 @@ a bowl full of gold bars sitting on a table
  - Prompt metadata saved into image file metadata
  - Edit images by describing the part you want edited (see example above)
  - Have AI generate captions for images `aimg describe <filename-or-url>`
+ - Interactive prompt: just run `aimg`
 
 ## How To
 
@@ -180,7 +181,7 @@ docker run -it --gpus all -v $HOME/.cache/huggingface:/root/.cache/huggingface -
 [Example Colab](https://colab.research.google.com/drive/1rOvQNs0Cmn_yU1bKWjCOHzGVDgZkaTtO?usp=sharing)
 
 ## ChangeLog
- 
+ - feature: interactive prompt added. access by running `aimg`
  - feature: Specify advanced text based masks using boolean logic and strength modifiers. Mask descriptions must be lowercase. Keywords uppercase.
    Valid symbols: `AND`, `OR`, `NOT`, `()`, and mask strength modifier `{+0.1}` where `+` can be any of `+ - * /`    
  - feature: apply mask edits to original files
@@ -285,7 +286,7 @@ docker run -it --gpus all -v $HOME/.cache/huggingface:/root/.cache/huggingface -
    - CPU support
    - ✅ img2img for plms
    - img2img for kdiff functions
-   - image masking
+   - text based image masking
      - https://boolean-parser.readthedocs.io/en/latest/index.html
      - https://github.com/facebookresearch/detectron2
    - images as actual prompts instead of just init images

imaginairy/api.py

@@ -91,7 +91,7 @@ def imagine_image_files(
     record_step_images=False,
     output_file_extension="jpg",
     print_caption=False,
-    create_modified_originals_for_masks=True,
+    mask_modify_original=True,
 ):
     big_path = os.path.join(outdir, "upscaled")
     masked_orig_path = os.path.join(outdir, "modified_originals")
@@ -120,7 +120,7 @@ def imagine_image_files(
         ddim_eta=ddim_eta,
         img_callback=_record_step if record_step_images else None,
         add_caption=print_caption,
-        create_modified_originals_for_masks=create_modified_originals_for_masks,
+        mask_modify_original=mask_modify_original,
     ):
         prompt = result.prompt
         basefilename = f"{base_count:06}_{prompt.seed}_{prompt.sampler_type}{prompt.steps}_PS{prompt.prompt_strength}_{prompt_normalized(prompt.prompt_text)}"
@@ -151,7 +151,7 @@ def imagine(
     img_callback=None,
     half_mode=None,
     add_caption=False,
-    create_modified_originals_for_masks=True,
+    mask_modify_original=True,
 ):
     model = load_model()
 
@@ -335,11 +335,7 @@ def imagine(
                             upscaled_img = enhance_faces(upscaled_img, fidelity=0.8)
 
                     # put the newly generated patch back into the original, full size image
-                    if (
-                        create_modified_originals_for_masks
-                        and mask_image_orig
-                        and prompt.init_image
-                    ):
+                    if mask_modify_original and mask_image_orig and prompt.init_image:
                         img_to_add_back_to_original = (
                             upscaled_img if upscaled_img else img
                         )

imaginairy/cmds.py

@@ -1,6 +1,7 @@
 import logging.config
 
 import click
+from click_shell import shell
 
 from imaginairy import LazyLoadingImage, generate_caption
 from imaginairy.api import imagine_image_files, load_model
@@ -111,12 +112,17 @@ def configure_logging(level="INFO"):
     type=click.Choice(["DEBUG", "INFO", "WARNING", "ERROR"]),
     help="What level of logs to show.",
 )
+@click.option(
+    "--quiet, -q",
+    is_flag=True,
+    type=click.Choice(["DEBUG", "INFO", "WARNING", "ERROR"]),
+    help="Alias of `--log-level ERROR`",
+)
 @click.option(
     "--show-work",
     default=["none"],
-    type=click.Choice(["none", "images", "video"]),
-    multiple=True,
-    help="Make a video showing the image being created",
+    is_flag=True,
+    help="Output a debug images to `steps` folder.",
 )
 @click.option(
     "--tile",
@@ -144,6 +150,12 @@ def configure_logging(level="INFO"):
     type=click.Choice(["keep", "replace"]),
     help="Should we replace the masked area or keep it?",
 )
+@click.option(
+    "--mask-modify-original",
+    default=True,
+    is_flag=True,
+    help="After the inpainting is done",
+)
 @click.option(
     "--caption",
     default=False,
@@ -174,6 +186,7 @@ def imagine_cmd(
     sampler_type,
     ddim_eta,
     log_level,
+    quiet,
     show_work,
     tile,
     mask_image,
@@ -186,6 +199,8 @@ def imagine_cmd(
     if ctx.invoked_subcommand is not None:
         return
     suppress_annoying_logs_and_warnings()
+    if quiet:
+        log_level = "ERROR"
     configure_logging(log_level)
 
     total_image_count = len(prompt_texts) * repeats
@@ -233,7 +248,7 @@ def imagine_cmd(
     )
 
 
-@click.group("aimg")
+@shell(prompt="aimg > ", intro="Starting imaginAIry...")
 def aimg():
     pass
 
@@ -253,7 +268,7 @@ def describe(image_filepaths):
         print(generate_caption(img.copy()))
 
 
-aimg.add_command(imagine_cmd, name="generate")
+aimg.add_command(imagine_cmd, name="imagine")
 
 if __name__ == "__main__":
     imagine_cmd()  # noqa

requirements-dev.txt

@@ -41,8 +41,11 @@ charset-normalizer==2.1.1
 click==8.1.3
     # via
     #   black
+    #   click-shell
     #   imaginAIry (setup.py)
     #   typer
+click-shell==2.1
+    # via imaginAIry (setup.py)
 contourpy==1.0.5
     # via matplotlib
 coverage==6.4.4
@@ -59,7 +62,7 @@ facexlib==0.2.5
     # via
     #   gfpgan
     #   realesrgan
-fairscale==0.4.9
+fairscale==0.4.10
     # via imaginAIry (setup.py)
 filelock==3.8.0
     # via
@@ -95,7 +98,7 @@ google-auth-oauthlib==0.4.6
     # via
     #   tb-nightly
     #   tensorboard
-grpcio==1.48.1
+grpcio==1.49.1
     # via
     #   tb-nightly
     #   tensorboard
@@ -316,7 +319,7 @@ six==1.16.0
     #   python-dateutil
 snowballstemmer==2.2.0
     # via pydocstyle
-tb-nightly==2.11.0a20220921
+tb-nightly==2.11.0a20220923
     # via
     #   basicsr
     #   gfpgan

setup.py

@@ -32,6 +32,7 @@ setup(
     },
     install_requires=[
         "click",
+        "click-shell",
         "protobuf != 3.20.2, != 3.19.5",
         "fairscale>=0.4.4",  # for vendored blip
         "ftfy",  # for vendored clip