feat: improve llm qualifier concurrency, rationalize args, standardize schema w/ pydantic

Author: percevalw

docs/pipes/llm/index.md

@@ -10,5 +10,6 @@ to perform various information extraction tasks.
 | Component                  | Description                                               |
 |----------------------------|-----------------------------------------------------------|
 | `eds.llm_markup_extractor` | Extract structured information using LLMs through markup. |
+| `eds.llm_span_qualifier`   | Predict attributes of spans using LLMs.                   |
 
 <!-- --8<-- [end:components] -->

docs/pipes/llm/llm-span-qualifier.md

@@ -0,0 +1,8 @@
+# LLM Span Qualifier {: #edsnlp.pipes.llm.llm_span_qualifier.factory.create_component }
+
+::: edsnlp.pipes.llm.llm_span_qualifier.factory.create_component
+    options:
+        heading_level: 3
+        show_bases: false
+        show_source: false
+        only_class_level: true

docs/pipes/qualifiers/llm-qualifier.md

@@ -84,4 +84,5 @@
     from .trainable.embeddings.text_cnn.factory import create_component as text_cnn
     from .misc.split import Split as split
     from .misc.explode import Explode as explode
-    from .llm.llm_markup_extractor import LlmMarkupExtractor as llm_markup_extractor
+    from .llm.llm_markup_extractor.factory import create_component as llm_markup_extractor
+    from .llm.llm_span_qualifier.factory import create_component as llm_span_qualifier

docs/tutorials/qualifying-entities-with-llm.md

@@ -0,0 +1 @@
+from .llm_span_qualifier import LlmSpanQualifier

edsnlp/pipes/__init__.py

@@ -1,7 +1,8 @@
-from edsnlp.core import registry
+from edsnlp import registry
 
-from .llm_span_qualifier import LLMSpanClassifier
+from .llm_span_qualifier import LlmSpanQualifier
 
 create_component = registry.factory.register(
     "eds.llm_span_qualifier",
-)(LLMSpanClassifier)
+    assigns=["doc.ents", "doc.spans"],
+)(LlmSpanQualifier)