add ref parser and summarizer

reference string parsing and summarization demos

app.py

@@ -0,0 +1,111 @@
+import gradio as gr
+from description import *
+
+from reference_string_parsing import *
+from summarization import *
+
+with gr.Blocks(css="#htext span {white-space: pre-line}") as demo:
+    gr.Markdown("# Gradio Demo for SciAssist")
+    with gr.Tabs():
+        # Reference String Parsing
+        with gr.TabItem("Reference String Parsing"):
+            with gr.Box():
+                gr.Markdown(rsp_str_md)
+                with gr.Row():
+                    with gr.Column():
+                        rsp_str = gr.Textbox(label="Input String")
+                        with gr.Column():
+                            rsp_str_dehyphen = gr.Checkbox(label="dehyphen")
+                        with gr.Row():
+                            rsp_str_btn = gr.Button("Parse")
+                    rsp_str_output = gr.HighlightedText(
+                        elem_id="htext",
+                        label="The Result of Parsing",
+                        combine_adjacent=True,
+                        adjacent_separator=" ",
+                    )
+                rsp_str_examples = gr.Examples(examples=[[
+                                                         "Waleed Ammar, Matthew E. Peters, Chandra Bhagavat- ula, and Russell Power. 2017. The ai2 system at semeval-2017 task 10 (scienceie): semi-supervised end-to-end entity and relation extraction. In ACL workshop (SemEval).",
+                                                         True],
+                                                     [
+                                                         "Isabelle Augenstein, Mrinal Das, Sebastian Riedel, Lakshmi Vikraman, and Andrew D. McCallum. 2017. Semeval-2017 task 10 (scienceie): Extracting keyphrases and relations from scientific publications. In ACL workshop (SemEval).",
+                                                         False]], inputs=[rsp_str, rsp_str_dehyphen])
+            with gr.Box():
+                gr.Markdown(rsp_file_md)
+                with gr.Row():
+                    with gr.Column():
+                        rsp_file = gr.File(label="Input File")
+                        rsp_file_dehyphen = gr.Checkbox(label="dehyphen")
+                        with gr.Row():
+                            rsp_file_btn = gr.Button("Parse")
+
+                    rsp_file_output = gr.HighlightedText(
+                        elem_id="htext",
+                        label="The Result of Parsing",
+                        combine_adjacent=True,
+                        adjacent_separator=" ",
+                    )
+                rsp_file_examples = gr.Examples(examples=[["examples/N18-3011_ref.txt", False],], inputs=[rsp_file, rsp_file_dehyphen])
+
+
+        rsp_file_btn.click(
+            fn=rsp_for_file,
+            inputs=[rsp_file, rsp_file_dehyphen],
+            outputs=rsp_file_output
+        )
+        rsp_str_btn.click(
+            fn=rsp_for_str,
+            inputs=[rsp_str, rsp_str_dehyphen],
+            outputs=rsp_str_output
+        )
+
+        # Single Document Summarization
+        with gr.TabItem("Single Document Summarization"):
+            with gr.Box():
+                gr.Markdown(ssum_str_md)
+                with gr.Row():
+                    with gr.Column():
+                        ssum_str = gr.Textbox(label="Input String")
+                        with gr.Column():
+                            ssum_str_beams = gr.Number(label="Number of beams for beam search", value=1, precision=0)
+                            ssum_str_sequences = gr.Number(label="Number of generated summaries", value=1, precision=0)
+                        with gr.Row():
+                            ssum_str_btn = gr.Button("Generate")
+                    ssum_str_output = gr.Textbox(
+                        elem_id="htext",
+                        label="Summary",
+                    )
+                ssum_str_examples = gr.Examples(examples=[[ssum_str_example, 1, 1], ],
+                                                inputs=[ssum_str, ssum_str_beams, ssum_str_sequences])
+            with gr.Box():
+                gr.Markdown(ssum_file_md)
+                with gr.Row():
+                    with gr.Column():
+                        ssum_file = gr.File(label="Input File")
+                        with gr.Column():
+                            ssum_file_beams = gr.Number(label="Number of beams for beam search", value=1, precision=0)
+                            ssum_file_sequences = gr.Number(label="Number of generated summaries", value=1, precision=0)
+                        with gr.Row():
+                            ssum_file_btn = gr.Button("Generate")
+                    ssum_file_output = gr.Textbox(
+                        elem_id="htext",
+                        label="Summary",
+                    )
+                ssum_file_examples = gr.Examples(examples=[["examples/BERT_body.txt", 10, 2],],
+                                                inputs=[ssum_file, ssum_file_beams, ssum_file_sequences])
+
+    ssum_file_btn.click(
+        fn=ssum_for_file,
+        inputs=[ssum_file, ssum_file_beams, ssum_file_sequences],
+        outputs=ssum_file_output
+    )
+    ssum_str_btn.click(
+        fn=ssum_for_str,
+        inputs=[ssum_str, ssum_str_beams, ssum_str_sequences],
+        outputs=ssum_str_output
+    )
+
+
+
+
+demo.launch()

description.py

@@ -0,0 +1,30 @@
+# Reference string parsing Markdown
+rsp_str_md = '''
+To **test on strings**, simply input one or more strings.
+'''
+
+rsp_file_md = '''
+To **test on a file**, the input can be:
+
+- A txt file which contains a reference string in each line.
+ 
+
+'''
+# - A pdf file which contains a whole scientific document without any processing (including title, author...).
+
+ssum_str_md = '''
+To **test on strings**, simply input a string.
+
+**Note**: The **number of beams** should be **divisible** by the **number of generated summaries** for group beam search.
+
+'''
+
+ssum_file_md = '''
+To **test on a file**, the input can be:
+
+- A txt file which contains the content to be summarized.
+
+**Note**: The **number of beams** should be **divisible** by the **number of generated summaries** for group beam search.
+
+'''
+# - A pdf file which contains a whole scientific document without any processing (including title, author...).

reference_string_parsing.py

@@ -0,0 +1,36 @@
+from typing import List, Tuple
+import torch
+from SciAssist import ReferenceStringParsing
+
+device = "gpu" if torch.cuda.is_available() else "cpu"
+rsp_pipeline = ReferenceStringParsing()
+
+
+def rsp_for_str(input, dehyphen=False) -> List[Tuple[str, str]]:
+    results = rsp_pipeline.predict(input, type="str", dehyphen=dehyphen)
+    output = []
+    for res in results:
+        for token, tag in zip(res["tokens"], res["tags"]):
+            output.append((token, tag))
+        output.append(("\n\n", None))
+    return output
+
+
+def rsp_for_file(input, dehyphen=False) -> List[Tuple[str, str]]:
+    if input == None:
+        return None
+    filename = input.name
+    # Identify the format of input and parse reference strings
+    if filename[-4:] == ".txt":
+        results = rsp_pipeline.predict(filename, type="txt", dehyphen=dehyphen)
+    # elif filename[-4:] == ".pdf":
+    #     results = rsp_pipeline.predict(filename, dehyphen=dehyphen)
+    else:
+        return [("File Format Error !", None)]
+    # Prepare for the input gradio.HighlightedText accepts.
+    output = []
+    for res in results:
+        for token, tag in zip(res["tokens"], res["tags"]):
+            output.append((token, tag))
+        output.append(("\n\n", None))
+    return output

requirements.txt

@@ -0,0 +1,2 @@
+torch==1.12.0
+SciAssist==0.0.18

summarization.py

@@ -0,0 +1,37 @@
+from typing import List, Tuple
+import torch
+from SciAssist import Summarization
+
+device = "gpu" if torch.cuda.is_available() else "cpu"
+ssum_pipeline = Summarization()
+
+
+def ssum_for_str(input, num_beams=1, num_return_sequences=1) -> List[Tuple[str, str]]:
+    results = ssum_pipeline.predict(input, type="str", num_beams=num_beams, num_return_sequences=num_return_sequences)
+
+    output = []
+    for res in results["summary"]:
+        output.append(f"{res}\n\n")
+    return "".join(output)
+
+
+def ssum_for_file(input, num_beams=1, num_return_sequences=1) -> List[Tuple[str, str]]:
+    if input == None:
+        return None
+    filename = input.name
+    # Identify the format of input and parse reference strings
+    if filename[-4:] == ".txt":
+        results = ssum_pipeline.predict(filename, type="txt", num_beams=num_beams,
+                                       num_return_sequences=num_return_sequences, save_results=False)
+    # elif filename[-4:] == ".pdf":
+    #     results = rsp_pipeline.predict(filename, num_beams=num_beams, num_return_sequences=num_return_sequences)
+    else:
+        return [("File Format Error !", None)]
+
+    output = []
+    for res in results["summary"]:
+        output.append(f"{res}\n\n")
+    return "".join(output)
+
+
+ssum_str_example = "Language model pre-training has been shown to be effective for improving many natural language processing tasks ( Dai and Le , 2015 ; Peters et al. , 2018a ; Radford et al. , 2018 ; Howard and Ruder , 2018 ) . These include sentence-level tasks such as natural language inference ( Bowman et al. , 2015 ; Williams et al. , 2018 ) and paraphrasing ( Dolan and Brockett , 2005 ) , which aim to predict the relationships between sentences by analyzing them holistically , as well as token-level tasks such as named entity recognition and question answering , where models are required to produce fine-grained output at the token level ( Tjong Kim Sang and De Meulder , 2003 ; Rajpurkar et al. , 2016 ) . There are two existing strategies for applying pre-trained language representations to downstream tasks : feature-based and fine-tuning . The feature-based approach , such as ELMo ( Peters et al. , 2018a ) , uses task-specific architectures that include the pre-trained representations as additional features . The fine-tuning approach , such as the Generative Pre-trained Transformer ( OpenAI GPT ) ( Radford et al. , 2018 ) , introduces minimal task-specific parameters , and is trained on the downstream tasks by simply fine-tuning all pretrained parameters . The two approaches share the same objective function during pre-training , where they use unidirectional language models to learn general language representations . We argue that current techniques restrict the power of the pre-trained representations , especially for the fine-tuning approaches . The major limitation is that standard language models are unidirectional , and this limits the choice of architectures that can be used during pre-training . For example , in OpenAI GPT , the authors use a left-toright architecture , where every token can only attend to previous tokens in the self-attention layers of the Transformer ( Vaswani et al. , 2017 ) . Such restrictions are sub-optimal for sentence-level tasks , and could be very harmful when applying finetuning based approaches to token-level tasks such as question answering , where it is crucial to incorporate context from both directions . In this paper , we improve the fine-tuning based approaches by proposing BERT : Bidirectional Encoder Representations from Transformers . BERT alleviates the previously mentioned unidirectionality constraint by using a `` masked language model '' ( MLM ) pre-training objective , inspired by the Cloze task ( Taylor , 1953 ) . The masked language model randomly masks some of the tokens from the input , and the objective is to predict the original vocabulary id of the masked arXiv:1810.04805v2 [ cs.CL ] 24 May 2019 word based only on its context . Unlike left-toright language model pre-training , the MLM objective enables the representation to fuse the left and the right context , which allows us to pretrain a deep bidirectional Transformer . In addition to the masked language model , we also use a `` next sentence prediction '' task that jointly pretrains text-pair representations . The contributions of our paper are as follows : • We demonstrate the importance of bidirectional pre-training for language representations . Unlike Radford et al . ( 2018 ) , which uses unidirectional language models for pre-training , BERT uses masked language models to enable pretrained deep bidirectional representations . This is also in contrast to Peters et al . ( 2018a ) , which uses a shallow concatenation of independently trained left-to-right and right-to-left LMs . • We show that pre-trained representations reduce the need for many heavily-engineered taskspecific architectures . BERT is the first finetuning based representation model that achieves state-of-the-art performance on a large suite of sentence-level and token-level tasks , outperforming many task-specific architectures . • BERT advances the state of the art for eleven NLP tasks . The code and pre-trained models are available at https : //github.com/ google-research/bert . "