Custom Prompts Example Config, Docs

README.md

@@ -18,14 +18,14 @@ In order to revise your manuscript, prompts must be provided to the AI model. Th
 - **Default prompts**: you can use the default prompts provided by the tool, in which case you don't need to do anything.
 - **Custom prompts**: you can define your own prompts to apply to specific files using YAML configuration files that you include with your manuscript.
 
-The default prompt, which should work for most manuscripts, is the following:
+If you wish to customize the prompts on a per-file basis, see [docs/custom-prompts.md](docs/custom-prompts.md) for more information.
 
-```
-Proofread the following paragraph that is part of a scientific manuscript.
-Keep all Markdown formatting, citations to other articles, mathematical expressions, and equations.
-```
+### Caveats
 
-If you wish to customize the prompts on a per-file basis, see [docs/custom-prompts.md](docs/custom-prompts.md) for more information.
+In the current implementation, the editor can only process one paragraph at a time.
+This limits the contextual information the LLM receives and thus the specificity of what it can check and fix.
+For example, in the Discussion section of a manuscript, the first paragraph should typically summarize the findings from the Results section, while the rest of the paragraphs should follow a different structure, but the AI editor can only judge each paragraph in the same way.
+We plan to reduce or remove this limitation in the future.
 
 ### Command line
 
@@ -48,7 +48,7 @@ For example, to change the temperature parameter of OpenAI models, you can expor
 Then, within the root directory of your Manubot-based manuscript, run the following commands (**IMPORTANT:** this will overwrite your original manuscript!):
 
 ```bash
-manubot ai-revision --content-directory content/
+manubot ai-revision --content-directory content/ --config-directory ci/
 ```
 
 The tool will revise each paragraph of your manuscript and write back the revised files in the same directory.
@@ -59,6 +59,7 @@ Before using the OpenAI API and incurring costs, you can run a test by using a d
 ```bash
 manubot ai-revision \
   --content-directory content/ \
+  --config-directory ci/ \
   --model-type DummyManuscriptRevisionModel \
   --model-kwargs add_paragraph_marks=True
 ```
@@ -88,8 +89,12 @@ from manubot_ai_editor.models import GPT3CompletionModel
 # create a manuscript editor object
 # here content_dir points to the "content" directory of the Manubot-based
 # manuscript, where Markdown files (*.md) are located
+# config_dir points to where CI-related configuration, including the AI
+# editor's configuration, is stored. it's optional, and if left out will
+# resort to defaults.
 me = ManuscriptEditor(
     content_dir="content",
+    config_dir="ci"
 )
 
 # create a model to revise the manuscript

docs/custom-prompts.md

@@ -8,14 +8,15 @@ There are two ways that you can use the custom prompts system:
 2. You can create the `ai-revision-prompts.yaml`, but only specify prompts and identifiers, which makes it suitable for sharing with others who have different names for their manuscripts' files.
 You would then specify a second file, `ai-revision-config.yaml`, that maps the prompt identifiers to the actual files in your manuscript.
 
-These files should be placed in the `content` directory alongside your manuscript markdown files.
+These files should be placed in the `ci` directory under your manubot root directory.
 
 See [Functionality Notes](#functionality-notes) later in this document for more information on how to write regular expressions and use placeholders in your prompts.
 
+
 ## Approach 1: Single file
 
 With this approach, you can define your prompts and how they map to your manuscript files in a single file.
-The single file should be named `ai-revision-prompts.yaml` and placed in the `content` folder.
+The single file should be named `ai-revision-prompts.yaml` and placed in the `ci` folder.
 
 The file would look something like the following:
 

libs/manubot_ai_editor/editor.py

@@ -97,7 +97,7 @@ def revise_and_write_paragraph(
         Arguments:
             paragraph: list of lines of the paragraph.
             section_name: name of the section the paragraph belongs to.
-            resolved_prompt: a prompt resolved via the ai_revision prompt config; None if unavailable
+            resolved_prompt: a prompt resolved via the ai-revision prompt config; None if unavailable
             revision_model: model to use for revision.
             outfile: file object to write the revised paragraph to.
 
@@ -267,7 +267,7 @@ def revise_file(
             output_dir (Path | str): path to the directory where the revised file will be written.
             revision_model (ManuscriptRevisionModel): model to use for revision.
             section_name (str, optional): Defaults to None. If so, it will be inferred from the filename.
-            resolved_prompt (str, optional): A prompt resolved via ai_revision prompt config files, which overrides any custom or section-derived prompts; None if unavailable.
+            resolved_prompt (str, optional): A prompt resolved via ai-revision prompt config files, which overrides any custom or section-derived prompts; None if unavailable.
         """
         input_filepath = self.content_dir / input_filename
         assert input_filepath.exists(), f"Input file {input_filepath} does not exist"
@@ -466,7 +466,7 @@ def revise_manuscript(
         for filename in sorted(self.content_dir.glob("*.md")):
             filename_section = self.get_section_from_filename(filename.name)
 
-            # use the ai_revision prompt config to attempt to resolve a prompt
+            # use the ai-revision prompt config to attempt to resolve a prompt
             resolved_prompt, _ = self.prompt_config.get_prompt_for_filename(
                 filename.name
             )
@@ -477,7 +477,7 @@ def revise_manuscript(
 
             # we do not process the file if all hold:
             # 1. it has no section *or* resolved prompt
-            # 2. we're unable to resolve it via ai_revision prompt configuration
+            # 2. we're unable to resolve it via ai-revision prompt configuration
             # 2. there is no custom prompt
             if (filename_section is None and resolved_prompt is None) and (
                 env_vars.CUSTOM_PROMPT not in os.environ

libs/manubot_ai_editor/models.py

@@ -26,7 +26,7 @@ def revise_paragraph(self, paragraph_text, section_name, resolved_prompt=None):
         Args:
             paragraph_text (str): text of the paragraph to revise.
             section_name (str): name of the section the paragraph belongs to.
-            resolved_prompt (str): prompt resolved via ai_revision config files, if available
+            resolved_prompt (str): prompt resolved via ai-revision config files, if available
 
         Returns:
             Revised paragraph text.
@@ -265,7 +265,7 @@ def get_prompt(
         Args:
             paragraph_text: text of the paragraph to revise.
             section_name: name of the section the paragraph belongs to.
-            resolved_prompt: prompt resolved via ai_revision config, if available
+            resolved_prompt: prompt resolved via ai-revision config, if available
 
         Returns:
             If self.endpoint != "edits", then returns a string with the prompt to be used by the model for the revision of the paragraph.
@@ -314,7 +314,7 @@ def get_prompt(
             #  a simple workaround is to remove {paragraph_text} from the prompt
             prompt = custom_prompt.format(**placeholders)
         elif resolved_prompt:
-            # use the resolved prompt from the ai_revision config files, if available
+            # use the resolved prompt from the ai-revision config files, if available
             # replace placeholders with their actual values
             prompt = resolved_prompt.format(**placeholders)
         elif section_name in ("abstract",):

libs/manubot_ai_editor/prompt_config.py

@@ -185,7 +185,7 @@ def get_prompt_for_filename(
         resolved_default_prompt = None
         if use_default and self.prompts is not None:
             resolved_default_prompt = self.prompts.get(
-                get_obj_path(self.config, ("files", "default_prompt"), missing="default"),
+                get_obj_path(self.config, ("files", "default_prompt")),
                 None
             )