first attempt to hf spaces

config/config.yaml

@@ -20,6 +20,9 @@ debug:
   is_debug: false
   force_rebuild_storage: false
 
+document_parsing:
+  enabled: false
+
 
 hydra:
   verbose: false

requirements.txt

@@ -35,7 +35,6 @@ cryptography==41.0.2
 cupy==12.1.0
 cycler==0.12.0
 cymem==2.0.7
-dataclasses==0.8
 dataclasses-json==0.5.7
 datasets==2.14.2
 debugpy==1.6.7
@@ -43,7 +42,6 @@ decorator==5.1.1
 defusedxml==0.7.1
 dill==0.3.7
 effdet==0.4.1
-en-core-web-trf==3.6.1
 entrypoints==0.4
 et-xmlfile==1.1.0
 exceptiongroup==1.1.2

src/demo.py

@@ -55,10 +55,11 @@ class App:
     def __init__(self, cfg : DictConfig) -> None:
         self.cfg = cfg
 
-        log.info("Loading: Document Loader")
-        self.loader = hydra.utils.instantiate(cfg.document_loader)
-        log.info("Loading: Text Splitter")
-        self.splitter = hydra.utils.instantiate(cfg.text_splitter)
+        if cfg.document_parsing.enabled:
+            log.info("Loading: Document Loader")
+            self.loader = hydra.utils.instantiate(cfg.document_loader)
+            log.info("Loading: Text Splitter")
+            self.splitter = hydra.utils.instantiate(cfg.text_splitter)
         log.info("Loading: Text Embedding Model")
         self.text_embedding_model = hydra.utils.instantiate(cfg.text_embedding)
         log.info("Loading: Vector Store")
@@ -73,6 +74,9 @@ class App:
         # if vector store exists, load it
         
         if not Path(self.cfg.storage_path.vector_store).exists() or self.cfg.debug.force_rebuild_storage:
+            if not self.cfg.document_parsing.enabled:
+                raise ValueError("Document parsing is not enabled. Cannot build storage from scratch")
+
             message = (
                 "Vector store not found at %s. Building storage from scratch"
                 if not self.cfg.debug.force_rebuild_storage

src/gradio.py

@@ -0,0 +1,17 @@
+import gradio as gr
+from hydra import compose, initialize
+from omegaconf import OmegaConf
+
+from demo import App
+
+def main():
+    with initialize(version_base=None, config_path="../config", job_name="gradio_app"):
+        cfg = compose(config_name="config", overrides=["document_parsing.enabled=False"])
+        
+        app = App(cfg)
+
+        webapp = gr.ChatInterface(fn=app.ask_chat, examples=["hello", "hola", "merhaba"], title="LLM4SciLit")
+        webapp.launch(share=True)
+
+if __name__ == "__main__":
+    main()