Merge branch 'main' of github.com:SapienzaNLP/relik into main

relik/cli/cli.py

@@ -216,10 +216,9 @@ def serve(
         annotation_type=annotation_type,
         host=host,
         port=port,
+        frontend=frontend
     )
 
-    if frontend:
-        serve_gradio()
 
 
 if __name__ == "__main__":

relik/inference/annotator.py

@@ -348,6 +348,37 @@ def __call__(
                 self.window_manager = WindowManager(
                     self.tokenizer, self.sentence_splitter
                 )
+            else:
+                if isinstance(self.sentence_splitter, WindowSentenceSplitter):
+                    if not isinstance(window_size, int):
+                        logger.warning(
+                            "With WindowSentenceSplitter the window_size must be an integer. "
+                            f"Using the default window size {self.window_manager.window_size}."
+                            f"If you want to change the window size to `sentence` or `none`, "
+                            f"please create a new Relik instance."
+                        )
+                        window_size = self.window_manager.window_size
+                        window_stride = self.window_manager.window_stride
+                if isinstance(self.sentence_splitter, SpacySentenceSplitter):
+                    if window_size != "sentence":
+                        logger.warning(
+                            "With SpacySentenceSplitter the window_size must be `sentence`. "
+                            f"Using the default window size {self.window_manager.window_size}."
+                            f"If you want to change the window size to an integer or `none`, "
+                            f"please create a new Relik instance."
+                        )
+                        window_size = "sentence"
+                        window_stride = None
+                if isinstance(self.sentence_splitter, BlankSentenceSplitter):
+                    if window_size != "none" or window_stride is not None:
+                        logger.warning(
+                            "With BlankSentenceSplitter the window_size must be `none`. "
+                            f"Using the default window size {self.window_manager.window_size}."
+                            f"If you want to change the window size to an integer or `sentence`, "
+                            f"please create a new Relik instance."
+                        )
+                        window_size = "none"
+                        window_stride = None
 
             # sanity check for window size and stride
             if (
@@ -520,6 +551,8 @@ def __call__(
         windows = windows + blank_windows
         windows.sort(key=lambda x: (x.doc_id, x.offset))
 
+        print(windows)
+
         # if there is no reader, just return the windows
         if self.reader is None:
             # normalize window candidates to be a list of lists, like when the reader is used
@@ -530,6 +563,7 @@ def __call__(
             merged_windows = self.window_manager.merge_windows(windows)
 
         # transform predictions into RelikOutput objects
+        print(merged_windows)
         output = []
         for w in merged_windows:
             span_labels = []

relik/inference/data/window/manager.py

@@ -5,6 +5,7 @@
 from relik.inference.data.splitters.blank_sentence_splitter import BlankSentenceSplitter
 from relik.inference.data.splitters.base_sentence_splitter import BaseSentenceSplitter
 from relik.inference.data.splitters.window_based_splitter import WindowSentenceSplitter
+from relik.inference.data.splitters.spacy_sentence_splitter import SpacySentenceSplitter
 from relik.inference.data.tokenizers.base_tokenizer import BaseTokenizer
 from relik.reader.data.relik_reader_sample import RelikReaderSample
 from relik.inference.data.objects import AnnotationType, TaskType
@@ -227,10 +228,10 @@ def _merge_doc_windows(self, windows: List[RelikReaderSample]) -> RelikReaderSam
         if len(windows) == 1:
             return self._normalize_single_window(windows[0])
 
-        if not isinstance(self.splitter, WindowSentenceSplitter):
-            # here we don't really need to merge windows, just normalize them
-            # TODO: check if we need to merge windows in this case
-            return [self._normalize_single_window(w) for w in windows]
+        # if not isinstance(self.splitter, WindowSentenceSplitter):
+        #     # here we don't really need to merge windows, just normalize them
+        #     # TODO: check if we need to merge windows in this case
+        #     return [self._normalize_single_window(w) for w in windows]
 
         if len(windows) > 0 and getattr(windows[0], "offset", None) is not None:
             windows = sorted(windows, key=(lambda x: x.offset))
@@ -507,7 +508,10 @@ def _merge_window_pair(
         ) = self._merge_predictions(window1, window2)
 
         # merge text, take into account overlapping chars
-        m_text = window1.text[: window2.offset] + window2.text
+        if isinstance(self.splitter, SpacySentenceSplitter):
+            m_text = window1.text[: window2.offset] + " "  + window2.text
+        else:
+            m_text = window1.text[: window2.offset] + window2.text
 
         merging_output.update(
             dict(

relik/inference/serve/backend/fastapi_be.py

@@ -148,6 +148,16 @@ async def relik_endpoint(
         relation_threshold: float = 0.5,
     ) -> List:
         try:
+            if window_size:
+                # check if window size is a number as string
+                if window_size.isdigit():
+                    window_size = int(window_size)
+
+            if window_stride:
+                # check if window stride is a number as string
+                if window_stride.isdigit():
+                    window_stride = int(window_stride)
+
             # get predictions for the retriever
             return await self(
                 text=text,
@@ -182,6 +192,7 @@ def main(
     workers: int = None,
     host: str = "localhost",
     port: int = 8000,
+    frontend: bool = False,
 ):
     app = FastAPI(
         title="ReLiK - A blazing fast and lightweight Information Extraction model for Entity Linking and Relation Extraction.",
@@ -201,6 +212,12 @@ def main(
         annotation_type=annotation_type,
     )
     app.include_router(server.router)
+    if frontend:
+        from relik.inference.serve.frontend.gradio_fe import main as serve_frontend
+        import threading
+
+        threading.Thread(target=serve_frontend, daemon=True).start()
+
     uvicorn.run(app, host=host, port=port, log_level="info", workers=workers)
 
 

relik/inference/serve/frontend/gradio_fe.py

@@ -305,22 +305,8 @@ def generate_graph(
 RELIK = os.getenv("RELIK", "localhost:8000/api/relik")
 
 
-def text_analysis(Text, Model, Relation_Threshold, Window_Size, Window_Stride):
-    global loaded_model
-    if Model is None:
-        return "", ""
-    # if loaded_model is None or loaded_model["key"] != Model:
-    #     relik = Relik.from_pretrained(Model, index_precision="bf16")
-    #     loaded_model = {"key": Model, "model": relik}
-    # else:
-    #     relik = loaded_model["model"]
-    # if Model not in relik_models:
-    #     raise ValueError(f"Model {Model} not found.")
-    # relik = relik_models[Model]
-    # spacy for span visualization
-
+def text_analysis(Text, Relation_Threshold, Window_Size, Window_Stride):
     relik = RELIK
-
     nlp = spacy.blank("xx")
     # annotated_text = relik(
     #     Text,
@@ -331,8 +317,10 @@ def text_analysis(Text, Model, Relation_Threshold, Window_Size, Window_Stride):
     #     window_size=Window_Size,
     #     window_stride=Window_Stride,
     # )
+    print(f"Using ReLiK at {relik}")
+    print(f"Querying ReLiK with ?text={Text}&relation_threshold={Relation_Threshold}&window_size={Window_Size}&window_stride={Window_Stride}&annotation_type=word&remove_nmes=False")
     response = requests.get(
-        f"{relik}?text={Text}&relation_threshold={Relation_Threshold}&window_size={Window_Size}&window_stride={Window_Stride}&annotation_type=word&remove_nmes=False"
+        f"http://{relik}/?text={Text}&relation_threshold={Relation_Threshold}&window_size={Window_Size}&window_stride={Window_Stride}&annotation_type=word&remove_nmes=False",
     )
     if response.status_code != 200:
         raise gr.Error(response.text)
@@ -390,11 +378,6 @@ def text_analysis(Text, Model, Relation_Threshold, Window_Size, Window_Stride):
         text_analysis,
         [
             gr.Textbox(label="Input Text", placeholder="Enter sentence here..."),
-            # gr.Dropdown(
-            #     relik_available_models,
-            #     value=relik_available_models[0],
-            #     label="Relik Model",
-            # ),
             gr.Slider(
                 minimum=0,
                 maximum=1,