import am anfang

app.py

@@ -25,6 +25,36 @@ import re
 import nltk
 nltk.download("stopwords", quiet=True)
 from nltk.corpus import stopwords as nltk_stopwords
+from nlp4web_codebase.ir.data_loaders.sciq import load_sciq
+
+
+import gradio as gr
+from typing import TypedDict
+
+
+from __future__ import annotations
+from dataclasses import asdict, dataclass
+import math
+import os
+from typing import Iterable, List, Optional, Type
+import tqdm
+from nlp4web_codebase.ir.data_loaders.dm import Document
+
+from nlp4web_codebase.ir.models import BaseRetriever
+from typing import Type
+from abc import abstractmethod
+
+
+from nlp4web_codebase.ir.data_loaders import Split
+import pytrec_eval
+import numpy as np
+
+
+from matplotlib import pyplot as plt
+
+from scipy.sparse._csc import csc_matrix
+
+
 
 LANGUAGE = "english"
 word_splitter = re.compile(r"(?u)\b\w\w+\b").findall
@@ -147,20 +177,11 @@ def run_counting(
         doc_texts=doc_texts,
     )
 
-from nlp4web_codebase.ir.data_loaders.sciq import load_sciq
 sciq = load_sciq()
 counting = run_counting(documents=iter(sciq.corpus), ndocs=len(sciq.corpus))
 
 """### BM25 Index"""
 
-from __future__ import annotations
-from dataclasses import asdict, dataclass
-import math
-import os
-from typing import Iterable, List, Optional, Type
-import tqdm
-from nlp4web_codebase.ir.data_loaders.dm import Document
-
 
 @dataclass
 class BM25Index(InvertedIndex):
@@ -257,9 +278,7 @@ bm25_index.save("output/bm25_index")
 
 """### BM25 Retriever"""
 
-from nlp4web_codebase.ir.models import BaseRetriever
-from typing import Type
-from abc import abstractmethod
+
 
 
 class BaseInvertedIndexRetriever(BaseRetriever):
@@ -330,10 +349,6 @@ Tune b and k1 on the **dev** split of SciQ using the metric MAP@10. The evaluati
 $${\displaystyle {\text{score}}(D,Q)=\sum _{i=1}^{n}{\text{IDF}}(q_{i})\cdot {\frac {f(q_{i},D)\cdot (k_{1}+1)}{f(q_{i},D)+k_{1}\cdot \left(1-b+b\cdot {\frac {|D|}{\text{avgdl}}}\right)}}}$$
 """
 
-from nlp4web_codebase.ir.data_loaders import Split
-import pytrec_eval
-import numpy as np
-
 
 
 def evaluate_map(rankings: Dict[str, Dict[str, float]], split=Split.dev) -> float:
@@ -346,7 +361,6 @@ def evaluate_map(rankings: Dict[str, Dict[str, float]], split=Split.dev) -> floa
 """Example of using the pre-requisite code:"""
 
 # Loading dataset:
-from nlp4web_codebase.ir.data_loaders.sciq import load_sciq
 sciq = load_sciq()
 counting = run_counting(documents=iter(sciq.corpus), ndocs=len(sciq.corpus))
 
@@ -444,7 +458,6 @@ print(plots_b["Y"][1])
 print(plots_k1)
 print(plots_b)
 
-from matplotlib import pyplot as plt
 plt.plot(plots_b["X"], plots_b["Y"], label="b")
 plt.plot(plots_k1["X"], plots_k1["Y"], label="k1")
 plt.ylabel("MAP")
@@ -486,7 +499,6 @@ Convert the matrix \begin{bmatrix}
 \end{bmatrix} to a `csc_matrix` by specifying `data`, `indices`, `indptr` and `shape`.
 """
 
-from scipy.sparse._csc import csc_matrix
 input_matrix = [[0, 1, 0, 3], [10, 2, 1, 0], [0, 0, 0, 9]]
 data = None
 indices = None
@@ -760,9 +772,6 @@ def search(query: str) -> List[Hit]:
 ```
 """
 
-import gradio as gr
-from typing import TypedDict
-
 class Hit(TypedDict):
   cid: str
   score: float