Update docstrings

.github/copilot-instructions.md

@@ -90,8 +90,11 @@ Use `pixi add --feature python-dev <package>` to add a dependency that is only u
 - Never include a period (".") at the end of the first line of docstrings.
 - Do not add a short summary to __init__ methods. Instead, keep the line blank and start the "Parameters" section after a second newline.
 - Do not document parameters in the class docstring - do that in the __init__ docstring instead.
-- All @property and @cached_property method documentation should be one line long and should start with the return type.
-- "Protected" functions and methods should always be documented using only one-line summary docstrings.
+- Do not add docstring to dunder methods (e.g., __str__, __repr__, etc.) unless absolutely necessary.
+- All @property and @cached_property method documentation should be one line long and should start with the return type followed by a colon (e.g. `"""str: My string property"""`).
+- If a parameter has a default value, always end the description with the sentence `"By default, <default value>."`
+- If the default value for a parameter is **not** `None`, document it using the format: `param_name : type, default=<default value>`. If the default value for a parameter **is** `None`, use the format : `param_name : type, optional`.
+- "Protected" functions and methods (i.e. starting with an underscore) should always be documented using **only** one-line summary docstrings.
 
 ## 7. Coding Guidelines (Rust)
 - Workspace-managed deps; update root `Cargo.toml` if adding shared dependency.

compass/_cli/main.py

@@ -11,7 +11,7 @@
 @click.version_option(version=__version__)
 @click.pass_context
 def main(ctx):
-    """Ordinance command line interface"""
+    """COMPASS command line interface"""
     ctx.ensure_object(dict)
 
 

compass/common/base.py

@@ -128,15 +128,53 @@ def llm_response_does_not_start_with_no(response):
 def setup_async_decision_tree(
     graph_setup_func, usage_sub_label=None, **kwargs
 ):
-    """Setup Async Decision tree for ordinance extraction"""
+    """Setup an ``AsyncDecisionTree`` for ordinance extraction
+
+    Parameters
+    ----------
+    graph_setup_func : callable
+        Factory that returns a fully configured
+        :class:`networkx.DiGraph`.
+    usage_sub_label : str, optional
+        Optional usage label reported to the LLM usage tracker.
+    **kwargs
+        Keyword arguments forwarded to ``graph_setup_func``.
+
+    Returns
+    -------
+    AsyncDecisionTree
+        Decision tree wrapping the graph produced by
+        ``graph_setup_func``.
+
+    Notes
+    -----
+    The function asserts that the tree has recorded at least the system
+    prompt before returning the constructed wrapper.
+    """
     G = graph_setup_func(**kwargs)  # noqa: N806
     tree = AsyncDecisionTree(G, usage_sub_label=usage_sub_label)
     assert len(tree.chat_llm_caller.messages) == 1
     return tree
 
 
 async def run_async_tree(tree, response_as_json=True):
-    """Run Async Decision Tree and return output as dict"""
+    """Run an async decision tree and optionally parse JSON output
+
+    Parameters
+    ----------
+    tree : AsyncDecisionTree
+        Decision tree to execute.
+    response_as_json : bool, default=True
+        If ``True``, attempts to parse the LLM response as JSON using
+        :func:`compass.utilities.parsing.llm_response_as_json`.
+        By default, ``True``.
+
+    Returns
+    -------
+    dict or str or None
+        Parsed dictionary when ``response_as_json`` is ``True``, raw
+        response otherwise. Returns ``None`` if execution fails.
+    """
     try:
         response = await tree.async_run()
     except COMPASSRuntimeError:
@@ -149,14 +187,39 @@ async def run_async_tree(tree, response_as_json=True):
 
 
 async def run_async_tree_with_bm(tree, base_messages):
-    """Run Async Decision Tree from base messages; return dict output"""
+    """Run an async decision tree using seed "base" messages
+
+    Parameters
+    ----------
+    tree : AsyncDecisionTree
+        Decision tree to execute.
+    base_messages : list of dict
+        Messages to preload into the tree's chat caller before running.
+
+    Returns
+    -------
+    dict or str or None
+        Output from :func:`run_async_tree`, filtered by the
+        ``response_as_json`` default.
+    """
     tree.chat_llm_caller.messages = base_messages
     assert len(tree.chat_llm_caller.messages) == len(base_messages)
     return await run_async_tree(tree)
 
 
 def empty_output(feature):
-    """Empty output for a feature (not found in text)"""
+    """Return the default empty result for a missing feature
+
+    Parameters
+    ----------
+    feature : str
+        Name of the feature to seed in the empty output structure.
+
+    Returns
+    -------
+    list of dict
+        Empty result placeholders used by downstream extraction logic.
+    """
     if feature in {"structures", "property line"}:
         return [
             {"feature": f"{feature} (participating)"},

compass/common/tree.py

@@ -66,32 +66,22 @@ def __init__(self, graph, usage_sub_label=None):
 
     @property
     def chat_llm_caller(self):
-        """ChatLLMCaller: ChatLLMCaller instance for this tree"""
+        """ChatLLMCaller: LLM caller bound to the decision tree"""
         return self.graph.graph["chat_llm_caller"]
 
     @cached_property
     def tree_name(self):
-        """str: Name of the decision tree"""
+        """str: Configured decision tree name"""
         return self._g.graph.get("_d_tree_name", "Unknown decision tree")
 
     @property
     def messages(self):
-        """Get a list of the conversation messages with the LLM
-
-        Returns
-        -------
-        list
-        """
+        """list: Conversation messages exchanged with the LLM"""
         return self.chat_llm_caller.messages
 
     @property
     def all_messages_txt(self):
-        """Get a printout of the full conversation with the LLM
-
-        Returns
-        -------
-        str
-        """
+        """str: Formatted conversation transcript"""
         messages = [
             f"{msg['role'].upper()}: {msg['content']}" for msg in self.messages
         ]
@@ -140,6 +130,12 @@ async def async_run(self, node0="init"):
         out : str or None
             Final response from LLM at the leaf node or ``None`` if an
             ``AttributeError`` was raised during execution.
+
+        Raises
+        ------
+        compass.exceptions.COMPASSRuntimeError
+            Raised when the traversal encounters an unexpected
+            exception that is not an ``AttributeError``.
         """
 
         self._history = []

compass/exceptions.py

@@ -10,7 +10,6 @@ class COMPASSError(Exception):
     """Generic COMPASS Error"""
 
     def __init__(self, *args, **kwargs):
-        """Init exception and broadcast message to logger"""
         super().__init__(*args, **kwargs)
         if args:
             logger.error(str(args[0]), stacklevel=2)

compass/extraction/apply.py

@@ -37,13 +37,20 @@ async def check_for_ordinance_info(
         has the ``"contains_ord_info"`` key, it will not be processed.
         To force a document to be processed by this function, remove
         that key from the documents ``attrs``.
+    model_config : compass.llm.config.LLMConfig
+        Configuration describing which LLM service, splitter, and call
+        parameters should be used for extraction.
+    heuristic : object
+        Domain-specific heuristic implementing a ``check`` method to
+        qualify text chunks for further processing.
     tech : str
         Technology of interest (e.g. "solar", "wind", etc). This is
         used to set up some document validation decision trees.
-    text_splitter : LCTextSplitter, optional
-        Optional Langchain text splitter (or subclass instance), or any
-        object that implements a `split_text` method. The method should
-        take text as input (str) and return a list of text chunks.
+    ordinance_text_collector_class : type
+        Collector class invoked to capture ordinance text chunks.
+    permitted_use_text_collector_class : type, optional
+        Collector class used to capture permitted-use districts text.
+        When ``None``, the permitted-use workflow is skipped.
     usage_tracker : UsageTracker, optional
         Optional tracker instance to monitor token usage during
         LLM calls. By default, ``None``.
@@ -61,6 +68,12 @@ async def check_for_ordinance_info(
         and an ``"ordinance_text"`` key containing the ordinance text
         snippet. Note that the snippet may contain other info as well,
         but should encapsulate all of the ordinance text.
+
+    Notes
+    -----
+    The function updates progress bar logging as chunks are processed
+    and sets ``contains_district_info`` when
+    ``permitted_use_text_collector_class`` is provided.
     """
     if "contains_ord_info" in doc.attrs:
         return doc
@@ -134,6 +147,9 @@ async def extract_date(doc, model_config, usage_tracker=None):
     ----------
     doc : elm.web.document.BaseDocument
         A document potentially containing date information.
+    model_config : compass.llm.config.LLMConfig
+        Configuration describing which LLM service, splitter, and call
+        parameters should be used for date extraction.
     usage_tracker : UsageTracker, optional
         Optional tracker instance to monitor token usage during
         LLM calls. By default, ``None``.
@@ -145,6 +161,11 @@ async def extract_date(doc, model_config, usage_tracker=None):
         the parsing are stored in the documents attrs. In particular,
         the attrs will contain a ``"date"`` key that will contain the
         parsed date information.
+
+    Notes
+    -----
+    Documents already containing a ``"date"`` attribute are returned
+    without reprocessing.
     """
     if "date" in doc.attrs:
         logger.debug(
@@ -186,8 +207,9 @@ async def extract_ordinance_text_with_llm(
         Optional Langchain text splitter (or subclass instance), or any
         object that implements a `split_text` method. The method should
         take text as input (str) and return a list of text chunks.
-    extractor : WindOrdinanceTextExtractor
-        Object used for ordinance text extraction.
+    extractor : object
+        Extractor instance exposing ``parsers`` that consume text
+        chunks and update ``doc.attrs``.
     original_text_key : str
         String corresponding to the `doc.attrs` key containing the
         original text (before extraction).
@@ -250,6 +272,9 @@ async def extract_ordinance_text_with_ngram_validation(
         Optional Langchain text splitter (or subclass instance), or any
         object that implements a `split_text` method. The method should
         take text as input (str) and return a list of text chunks.
+    extractor : object
+        Extractor instance exposing ``parsers`` that consume text
+        chunks and update ``doc.attrs``.
     original_text_key : str
         String corresponding to the `doc.attrs` key containing the
         original text (before extraction).
@@ -406,6 +431,9 @@ async def extract_ordinance_values(doc, parser, text_key, out_key):
         that are found to contain ordinance data. Note that if the
         document's attrs does not contain the `text_key` key, it will
         not be processed.
+    parser : object
+        Parser instance with an async ``parse`` method that converts
+        cleaned ordinance text into structured values.
     text_key : str
         Name of the key under which cleaned text is stored in
         `doc.attrs`. This text should be ready for extraction.
@@ -418,6 +446,11 @@ async def extract_ordinance_values(doc, parser, text_key, out_key):
     elm.web.document.BaseDocument
         Document that has been parsed for ordinance values. The results
         of the extraction are stored in the document's attrs.
+
+    Notes
+    -----
+    When the cleaned text is missing or empty the function emits a
+    :class:`compass.warn.COMPASSWarning` and leaves ``doc`` unchanged.
     """
     if not doc.attrs.get(text_key):
         msg = (