use nested attrs instead of dict items

README.md

@@ -170,7 +170,7 @@ for dump, _ in wiki.mediawikis.items():
     # Modify the source metadata by accounting for duplicates (<name>_n)
     # And add the mediawiki title (<name>_n - <wikiname>)
     merged_documents.extend(
-        Document(doc.page_content, {"source": doc.metadata["source"] + f" - {wiki}"})
+        Document(doc.page_content, {"source": doc.metadata["source"] + f" - {dump}"})
         for doc in rename_duplicates(loader.load())
     )
     ### Insert a new loader

main.py

@@ -1,6 +1,6 @@
+from collections import namedtuple
 from chainlit.input_widget import Slider, TextInput
 from chainlit.playground.config import add_llm_provider
-from collections import namedtuple
 from langchain.cache import SQLiteCache
 from langchain.callbacks.manager import CallbackManager
 from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
@@ -19,6 +19,7 @@
 import yaml
 
 import torch
+
 if not torch.cuda.is_available():
     torch.set_num_threads(torch.get_num_threads() * 2)
 
@@ -34,23 +35,30 @@ def __init__(self):
         except yaml.YAMLError as e:
             print(f"Error reading YAML file: {e}")
             exit(1)
+        self.convert_struct(**data)
 
-        for key, val in data.items():
-            if key == "mediawikis":
-                self.mediawikis = {wiki: "" for wiki in data["mediawikis"]}
+    def __getattr__(self, attr):
+        return self[attr]
+
+    def convert_struct(self, **kwargs):
+        for key, value in kwargs.items():
+            if isinstance(value, dict):
+                self.__dict__[key] = self.convert_struct(**value)
             else:
-                setattr(self, key, val)
+                self.__dict__[key] = value
+        return self
 
     def set_args(self, args):
         self.args = args
 
     def set_chat_settings(self, settings):
         # Update global wiki, not local self
         global wiki
-        if not settings:
-            wiki.inputs = wiki.settings
-        else:
-            wiki.inputs = settings
+        if isinstance(settings, dict):
+            for key, val in settings.items():
+                setattr(wiki, key, val)
+        if settings:
+            wiki.settings = settings
 
 
 ### Globals
@@ -77,7 +85,7 @@ def create_vector_db():
     Document = namedtuple("Document", ["page_content", "metadata"])
     merged_documents = []
 
-    for dump, _ in wiki.mediawikis.items():
+    for dump in wiki.mediawikis:
         # https://python.langchain.com/docs/integrations/document_loaders/mediawikidump
         loader = MWDumpLoader(
             encoding="utf-8",
@@ -91,7 +99,9 @@ def create_vector_db():
         # Modify the source metadata by accounting for duplicates (<name>_n)
         # And add the mediawiki title (<name>_n - <wikiname>)
         merged_documents.extend(
-            Document(doc.page_content, {"source": doc.metadata["source"] + f" - {dump}"})
+            Document(
+                doc.page_content, {"source": doc.metadata["source"] + f" - {dump}"}
+            )
             for doc in rename_duplicates(loader.load())
         )
     print(f"Embedding {len(merged_documents)} Pages, this may take a while.")
@@ -128,26 +138,21 @@ def create_chain():
     # https://python.langchain.com/docs/integrations/llms/llm_caching
     set_llm_cache(SQLiteCache(database_path="memory/cache.db"))
     # https://python.langchain.com/docs/integrations/llms/ollama
-    # Wiki loses its global scope when used as a parameter in ChatOllama()
-    # Unsure as to why
-    inputs = wiki.inputs
     model = ChatOllama(
         cache=True,
         callback_manager=callback_manager,
         model=wiki.model,
-        repeat_penalty=inputs["repeat_penalty"],
-        temperature=inputs["temperature"],
-        top_k=inputs["top_k"],
-        top_p=inputs["top_p"],
+        repeat_penalty=wiki.repeat_penalty,
+        temperature=wiki.temperature,
+        top_k=wiki.top_k,
+        top_p=wiki.top_p,
     )
     # https://api.python.langchain.com/en/latest/chains/langchain.chains.conversational_retrieval.base.ConversationalRetrievalChain.html
     chain = ConversationalRetrievalChain.from_llm(
         chain_type="stuff",
         llm=model,
         memory=memory,
-        retriever=vectordb.as_retriever(
-            search_kwargs={"k": int(wiki.inputs["num_sources"])}
-        ),
+        retriever=vectordb.as_retriever(search_kwargs={"k": int(wiki.num_sources)}),
         return_source_documents=True,
     )
 
@@ -164,13 +169,13 @@ async def on_chat_start():
         TextInput(
             id="num_sources",
             label="# of Sources",
-            initial=str(wiki.inputs["num_sources"]),
+            initial=str(wiki.num_sources),
             description="Number of sources returned based on their similarity source. The same source can be returned more than once. (Default: 4)",
         ),
         Slider(
             id="temperature",
             label="Temperature",
-            initial=wiki.inputs["temperature"],
+            initial=wiki.temperature,
             min=0,
             max=1,
             step=0.1,
@@ -179,7 +184,7 @@ async def on_chat_start():
         Slider(
             id="repeat_penalty",
             label="Repeat Penalty",
-            initial=wiki.inputs["repeat_penalty"],
+            initial=wiki.repeat_penalty,
             min=0.5,
             max=2.5,
             step=0.1,
@@ -188,7 +193,7 @@ async def on_chat_start():
         Slider(
             id="top_k",
             label="Top K",
-            initial=wiki.inputs["top_k"],
+            initial=wiki.top_k,
             min=0,
             max=100,
             step=1,
@@ -197,7 +202,7 @@ async def on_chat_start():
         Slider(
             id="top_p",
             label="Top P",
-            initial=wiki.inputs["top_p"],
+            initial=wiki.top_p,
             min=0,
             max=1,
             step=0.1,
@@ -214,7 +219,7 @@ async def on_chat_start():
             is_chat=True,
             # Not enough context to LangchainGenericProvider
             # https://github.com/Chainlit/chainlit/blob/main/backend/chainlit/playground/providers/langchain.py#L27
-            inputs=wiki.inputs,
+            inputs=wiki.settings,
         )
     )
     await cl.ChatSettings(inputs).send()

test/test.py

@@ -16,19 +16,19 @@ def test_multiwiki():
     assert wiki.model == "volo"
     assert wiki.question == "How many eyestalks does a Beholder have?"
     assert wiki.source == "./sources"
-    assert wiki.mediawikis == {
-        # "dnd4e": "",
-        "dnd5e": "",
-        # "darksun": "",
-        "dragonlance": "",
-        "eberron": "",
-        # "exandria": "",
-        "greyhawk": "",
-        "forgottenrealms": "",
-        # "planescape": "",
-        # "ravenloft": "",
-        # "spelljammer": "",
-    }
+    assert wiki.mediawikis == [
+        # "dnd4e",
+        "dnd5e",
+        # "darksun",
+        "dragonlance",
+        "eberron",
+        # "exandria",
+        "forgottenrealms",
+        "greyhawk",
+        # "planescape",
+        # "ravenloft",
+        # "spelljammer",
+    ]
 
 
 def test_multiwiki_set_args():