Rename Pipe to Thread

README.md

@@ -150,7 +150,7 @@ uv run pytest -v -m "not apikey"
 ```
 databao/
   api.py                 # public entry: new_agent(...)
-  core/                  # Agent, Pipe, Executor, Visualizer abstractions
+  core/                  # Agent, Thread, Executor, Visualizer abstractions
   agents/                # Lighthouse (default) and React-DuckDB agents
   duckdb/                # DuckDB integration and tools
   visualizers/           # Vega-Lite chat visualizer and utilities

databao/__init__.py

@@ -8,15 +8,15 @@
 
 from databao.api import new_agent
 from databao.configs.llm import LLMConfig
-from databao.core import Agent, ExecutionResult, Executor, Opa, Pipe, VisualisationResult, Visualizer
+from databao.core import Agent, ExecutionResult, Executor, Opa, Thread, VisualisationResult, Visualizer
 
 __all__ = [
     "Agent",
     "ExecutionResult",
     "Executor",
     "LLMConfig",
     "Opa",
-    "Pipe",
+    "Thread",
     "VisualisationResult",
     "Visualizer",
     "__version__",

databao/core/__init__.py

@@ -2,7 +2,7 @@
 from databao.core.cache import Cache
 from databao.core.executor import ExecutionResult, Executor
 from databao.core.opa import Opa
-from databao.core.pipe import Pipe
+from databao.core.thread import Thread
 from databao.core.visualizer import VisualisationResult, Visualizer
 
-__all__ = ["Agent", "Cache", "ExecutionResult", "Executor", "Opa", "Pipe", "VisualisationResult", "Visualizer"]
+__all__ = ["Agent", "Cache", "ExecutionResult", "Executor", "Opa", "Thread", "VisualisationResult", "Visualizer"]

databao/core/agent.py

@@ -7,7 +7,7 @@
 from sqlalchemy import Connection, Engine
 
 from databao.configs.llm import LLMConfig
-from databao.core.pipe import Pipe
+from databao.core.thread import Thread
 
 if TYPE_CHECKING:
     from databao.core.cache import Cache
@@ -50,7 +50,7 @@ def __init__(
         self.__visualizer = visualizer
         self.__cache = cache
 
-        # Pipe/thread defaults
+        # Thread defaults
         self.__rows_limit = rows_limit
         self.__lazy_threads = lazy_threads
         self.__auto_output_modality = auto_output_modality
@@ -134,11 +134,11 @@ def thread(
         stream_plot: bool | None = None,
         lazy: bool | None = None,
         auto_output_modality: bool | None = None,
-    ) -> Pipe:
+    ) -> Thread:
         """Start a new thread in this agent."""
         if not self.__dbs and not self.__dfs:
             raise ValueError("No databases or dataframes registered in this agent.")
-        return Pipe(
+        return Thread(
             self,
             rows_limit=self.__rows_limit,
             stream_ask=stream_ask if stream_ask is not None else self.__stream_ask,

databao/core/executor.py

@@ -23,7 +23,7 @@ class OutputModalityHints(BaseModel):
     # Currently, the only modality that makes sense to request outside the Executor is visualization.
     # If Executor was responsible for plotting as well (instead of Visualizer), then we could fully control and
     # customize rendering in ExecutionResult._repr_mimebundle_.
-    # But now we need hints to tell Pipe how to handle plotting.
+    # But now we need hints to tell Thread how to handle plotting.
 
     should_visualize: bool = False
     """Whether the execution results can be visualized."""

databao/core/pipe.py → databao/core/thread.py

@@ -12,11 +12,11 @@
     from databao.core.visualizer import VisualisationResult
 
 
-class Pipe:
+class Thread:
     """A single conversational thread within an agent.
 
-    - Maintains its own message history (isolated from other pipes).
-    - Materializes data and visualizations eagerly or lazily and caches results per pipe.
+    - Maintains its own message history (isolated from other threads).
+    - Materializes data and visualizations eagerly or lazily and caches results per thread.
     - Exposes helpers to get the latest dataframe/text/plot/code.
     """
 
@@ -53,7 +53,6 @@ def __init__(
         self._visualization_result: VisualisationResult | None = None
         self._visualization_request: str | None = None
 
-        # N.B. Pipes/Threads are currently append-only and cannot be "forked".
         self._opas_processed_count = 0
         self._opas: list[Opa] = []
         self._meta: dict[str, Any] = {}
@@ -119,7 +118,7 @@ def code(self) -> str | None:
         return self._materialize_data(self._data_materialized_rows).code
 
     def meta(self) -> dict[str, Any]:
-        """Aggregated metadata from executor/visualizer for this pipe."""
+        """Aggregated metadata from executor/visualizer for this thread."""
         self._materialize_data(self._data_materialized_rows)
         return self._meta
 
@@ -146,9 +145,9 @@ def plot(
         return self._materialize_visualization(request, rows_limit if rows_limit else self._data_materialized_rows)
 
     def ask(self, query: str, *, rows_limit: int | None = None, stream: bool | None = None) -> Self:
-        """Append a new user query to this pipe.
+        """Append a new user query to this thread.
 
-        Returns self to allow chaining (e.g., pipe.ask("...")).
+        Returns self to allow chaining (e.g., thread.ask("...")).
 
         Setting rows_limit has no effect in lazy mode.
         """