Resolve pull merge conflict

README.md

@@ -4,7 +4,8 @@ emoji: 🏢
 colorFrom: blue
 colorTo: indigo
 sdk: gradio
-app_file: app.py 
+sdk_version: 3.0.22
+app_file: app.py
 pinned: false
 license: mit
 ---

requirements.txt

@@ -1,4 +1,4 @@
-gradio
+# gradio
 install
 transformers
 sentencepiece
@@ -6,6 +6,7 @@ sklearn
 git+https://github.com/ffreemt/align-model-pool
 sentence-transformers
 logzero
+set_loglevel
 icecream
 alive-progress
 more_itertools

ubee/__main__.py

@@ -1,4 +1,21 @@
-"""Gen ubee main."""
+"""Gen ubee main.
+
+private
+url = 'https://hf.space/embed/mikeee/zero-shot/+/api/predict'
+resp = httpx.post(
+    url,
+    json={"data": ["love", ",".join(["liebe", "this is test", "hate you"]), False]},
+    timeout=httpx.Timeout(None, connect=3),
+)
+resp.json()
+{'data': [{'label': 'liebe',
+   'confidences': [{'label': 'liebe', 'confidence': 0.8688847422599792},
+    {'label': 'this is test', 'confidence': 0.12558135390281677},
+    {'label': 'hate you', 'confidence': 0.005533925257623196}]}],
+ 'duration': 0.265749454498291,
+ 'average_duration': 4.639325571060181}
+
+"""
 # pylint: disable=unused-import, wrong-import-position, wrong-import-order, too-many-locals, broad-except, line-too-long
 
 import sys
@@ -14,6 +31,8 @@ import pandas as pd
 from icecream import ic
 from icecream import install as ic_install
 from logzero import logger
+from set_loglevel import set_loglevel
+logzero.loglevel(set_loglevel())
 
 # for embeddable python
 # if "." not in sys.path: sys.path.insert(0, ".")
@@ -22,6 +41,8 @@ from ubee import __version__
 from ubee.ubee import ubee
 
 # logzero.loglevel(10)
+logger.debug(" debug on ")
+
 ic_install()
 ic.configureOutput(
     includeContext=True,
@@ -30,17 +51,44 @@ ic.configureOutput(
 ic.enable()
 # ic.disenable()  # to turn off
 
+ic(" ic.enabled ")
+
+_ = """
+ic("Testing...")
+import model_pool
+from model_pool import fetch_check_aux
+print("model-pool version", model_pool.__version__)
+print("gradio version", gr.__version__)
+
+try:
+    fetch_check_aux.fetch_check_aux()
+except Exception as _:
+    ic(["fetch_check_aux.fetch_check_aux", _])
+
+from model_pool.load_model import load_model
+try:
+    clas = load_model("clas-l-user")
+except Exception as _:
+    ic(["load_model(\"clas-l-user\")", _])
+# """
+
+# _ = clas("love", ["liebe", "hate you", "test"])
+# print(_)
+# raise SystemExit("Exit by intention")
+# {'sequence': 'love', 'labels': ['liebe', 'test', 'hate you'],
+# 'scores': [0.8885253667831421, 0.10581762343645096, 0.005657028406858444]}
+# Runs OK
 
-def greet1(name):
-    """Dummy."""
-    return "Hello " + name + "!!"
+# text1 = ""
+# text2 = ""
+# thresh: float = 0.4
 
 
+# segment: str
 def ifn(
     text1,
     text2,
-    # segment: str
-    thresh: float
+    thresh
 ) -> Tuple[pd.DataFrame, pd.DataFrame, pd.DataFrame]:
     """Take inputs, return outputs.
 
@@ -60,6 +108,7 @@ def ifn(
     # return _
 
     res1_, res2_ = ubee(res1, res2, thresh)
+    # res1_, res2_ = res1, res2
 
     out_df = pd.DataFrame(
         zip_longest(res1, res2),
@@ -72,11 +121,17 @@ def ifn(
         _ = None
 
     # return out_df, pd.DataFrame(res1_, columns=["text1", "text2", "likelihood"]), _
+<<<<<<< HEAD
     return pd.DataFrame(res1_, columns=["text1", "text2", "likelihood"]).to_html(), _.to_html()
+=======
+    return pd.DataFrame(res1_, columns=["text1", "text2", "likelihood"]), _
+>>>>>>> 1b2837ac7145a0ce5ead7088aaa41165a6615ad4
 
 
 def main():
     """Create main entry."""
+    # global text1, text2, threash
+
     text_zh = Path("data/test_zh.txt").read_text(encoding="utf8")
     text_zh = [elm.strip() for elm in text_zh.splitlines() if elm.strip()][:10]
     text_zh = "\n\n".join(text_zh)
@@ -136,73 +191,6 @@ def main():
     lines = 15
     placeholder = "Type or paste text here"
 
-    inputs = [
-        gr.inputs.Textbox(
-            lines=lines, placeholder=placeholder, default=ex2_zh, label="text1"
-        ),
-        gr.inputs.Textbox(
-            lines=lines, placeholder=placeholder, default=ex2_en, label="text2"
-        ),
-        gr.inputs.Slider(
-            minimum=0.0,
-            maximum=1.0,
-            step=0.1,
-            default=0.4,
-            label="threshold",
-        ),
-    ]
-
-    out_df = gr.outputs.Dataframe(
-        headers=None,
-        max_rows=lines,  # 20
-        max_cols=None,
-        overflow_row_behaviour="paginate",
-        type="auto",
-        label="To be aligned",
-    )
-    aligned = gr.outputs.Dataframe(
-        headers=None,
-        max_rows=lines,  # 20
-        max_cols=None,
-        overflow_row_behaviour="paginate",
-        type="auto",
-        label="Aligned",
-    )
-    leftover = gr.outputs.Dataframe(
-        headers=None,
-        max_rows=lines,  # 20
-        max_cols=None,
-        overflow_row_behaviour="paginate",
-        type="auto",
-        label="Leftover",
-    )
-    outputs = [  # tot. 3
-        out_df,
-        aligned,
-        leftover,
-    ]
-
-    _ = """ # switch to blocks
-    iface = gr.Interface(
-        fn=greet,
-        # fn=ubee,
-        title=title,
-        theme=theme,
-        layout="vertical",  # horizontal unaligned
-        description=description,
-        article=article,
-        # inputs="text",
-        # outputs="text",
-        inputs=inputs,  # text1, text2, segment, thresh
-        outputs=outputs,
-        examples=examples,
-        # enable_queue=True,
-    )
-    iface.launch(
-        enable_queue=True,
-        share=True,
-    )
-    # """
     blocks = gr.Blocks()
 
     with blocks:
@@ -212,27 +200,49 @@ def main():
 
             Align non-sequential dualtexts.
 
+            可对词、句、段，每个词（或句或段）一行。可对任意语言对（英中、英德、德法、中日……等等）。建议 threshold 门槛值 -- 词： 0.3，句：0.5， 段： 0.7。如果太多 leftover，可适当调小 threshold。 如果太多误对则可以适当调大 threshold。
+
             """).strip()
         )
         with gr.Column():
             with gr.Row():
+<<<<<<< HEAD
                 text1 = gr.inputs.Textbox(
                     lines=lines, placeholder=placeholder, default=ex1_zh, label="text1"
                 ),
                 text2 = gr.inputs.Textbox(
                     lines=lines, placeholder=placeholder, default=ex1_en, label="text2"
+=======
+                text1 = gr.Textbox(
+                    lines=lines,
+                    placeholder=placeholder,
+                    value=ex2_zh,
+                    label="text1"
+                )
+                text2 = gr.Textbox(
+                    lines=lines,
+                    placeholder=placeholder,
+                    value=ex2_en,
+                    label="text2"
+>>>>>>> 1b2837ac7145a0ce5ead7088aaa41165a6615ad4
                 )
             with gr.Row():
-                slider = gr.inputs.Slider(
+                thresh = gr.Slider(
                     minimum=0.0,
                     maximum=1.0,
                     step=0.1,
-                    default=0.4,
+                    value=0.4,
                     label="threshold",
                 )
                 btn = gr.Button("Run")
+<<<<<<< HEAD
             _ = """
             out_df = gr.outputs.Dataframe(
+=======
+
+            _ = """
+            out_df = gr.Dataframe(
+>>>>>>> 1b2837ac7145a0ce5ead7088aaa41165a6615ad4
                 headers=None,
                 max_rows=lines,  # 20
                 max_cols=None,
@@ -241,6 +251,7 @@ def main():
                 label="To be aligned",
             )
             # """
+<<<<<<< HEAD
             with gr.Row():
                 _ = """
                 aligned = gr.Dataframe(
@@ -262,21 +273,51 @@ def main():
                 # """
                 aligned = gr.HTML(value="output aligned", label="Aligned")
                 leftover = gr.HTML(value="output leftover", label="Leftover")
+=======
+
+            # with gr.Row():
+            aligned = gr.Dataframe(
+                headers=None,
+                max_rows=lines,  # 20
+                max_cols=None,
+                overflow_row_behaviour="paginate",
+                type="auto",
+                label="Aligned",
+            )
+
+            leftover = gr.Dataframe(
+                headers=None,
+                max_rows=lines,  # 20
+                max_cols=None,
+                overflow_row_behaviour="paginate",
+                type="auto",
+                label="Leftover",
+            )
+
+            logger.debug("text1: %s", text1)
+            logger.debug("text2: %s", text2)
+
+>>>>>>> 1b2837ac7145a0ce5ead7088aaa41165a6615ad4
             btn.click(
                 fn=ifn,
                 inputs=[
                     text1,
                     text2,
+                    thresh,
                 ],
                 outputs=[
-                    out_df,
+                    # out_df,
                     aligned,
                     leftover,
                 ]
             )
 
+    # blocks.launch()
+    blocks.launch(debug=True, enable_queue=True)
+
 
 if __name__ == "__main__":
+    logger.info(" Start main()")
     main()
 
 _ = """

ubee/uclas.py

@@ -13,9 +13,14 @@ from model_pool.load_model import load_model
 from model_pool.model_s import load_model_s
 from sklearn.metrics.pairwise import cosine_similarity
 
-logzero.loglevel(20)
+# logzero.loglevel(20)
+
+# fetch_check_aux("/home/user")
+try:
+    fetch_check_aux()
+except Exception as _:
+    logger.error(_)
 
-fetch_check_aux("/home/user")
 model_s = load_model_s()
 clas = load_model("clas-l-user")