1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
|
"""添加单词到本地生词本。"""
"""
add-word.py
用法：python3 add-word.py <单词>
"""
import sys
import os
import re
import requests
try:
from nltk.corpus import wordnet as wn
except ImportError:
wn = None
OLLAMA_MODEL = os.environ.get("OLLAMA_MODEL", "llama3.1:8b")
OLLAMA_BASE = os.environ.get("OLLAMA_BASE", "http://127.0.0.1:11434")
TARGET_FILE = os.path.join(os.path.dirname(__file__), "生单词.md")
def lookup_wordnet(word: str) -> dict:
"""直接调用 NLTK WordNet 查询，不依赖本地 JSON。"""
if wn is None:
return {}
try:
synsets = wn.synsets(word)
except Exception as e:
print(f"WordNet 查询失败：{e}")
return {}
meanings = []
examples = []
synonyms = set()
antonyms = set()
for s in synsets:
definition = (s.definition() or "").strip()
if definition:
meanings.append(definition)
for ex in s.examples():
ex = ex.strip()
if ex and ex not in examples:
examples.append(ex)
for lemma in s.lemma_names():
synonyms.add(lemma)
try:
for lemma in s.lemmas():
for ant in lemma.antonyms():
antonyms.add(ant.name())
except Exception:
pass
synonyms.discard(word)
synonyms.discard(word.lower())
synonyms.discard(word.capitalize())
return {
"meanings": meanings,
"examples": examples,
"synonyms": sorted(synonyms)[:10],
"antonyms": sorted(antonyms)[:10],
}
def ask_ollama(word: str, wordnet_data: dict) -> str:
"""调用本地 Ollama 生成词典条目。"""
meanings = wordnet_data.get("meanings", []) if wordnet_data else []
expected_count = len(meanings) if meanings else 5
prompt = (
f"你是一位英语词典助手。只输出词典内容，不要写代码、不要解释推理过程、不要输出多余文字。\n\n"
f"请针对单词 '{word}' 输出以下内容：\n"
f"1. 中文意思：这是最重要的一节，必须先完整输出。"
f"每条以词性缩写开头，如 adj. / n. / v. / adv.；"
f"必须严格输出 {expected_count} 条，一条对应一个义项，禁止把多个义项用分号合并到同一行；\n"
f"若某义项属于专业领域，请在中文后注明，例如「n.（产科）胎位/先露」；\n"
f"2. 用法：分两类——\n"
f" - 固定搭配：相对固定的短语，必须包含单词 '{word}' 本身；\n"
f" - 常见用法：一般性搭配或用法说明；\n"
f"3. 其他形式：只列该词的派生词、词形变化或同根词，例如 presenter / presentable / presentational；"
f"不要写近义词、不要写 unrelated words；格式为「单词（词性. 中文含义）」；\n"
f"4. 例句：为上面每一个中文义项各写 1 个英文例句，共 {expected_count} 条；"
f"每条例句中必须出现单词 '{word}' 本身，不得用它的动词/形容词/其他词形替代；"
f"每条后面用中文括号附翻译，例句要体现对应义项，不要重复。\n\n"
f"参考信息（来自 WordNet，可作辅助，不必照搬）：\n"
)
if meanings:
prompt += "释义（请逐条对应，不要遗漏）：\n"
for i, m in enumerate(meanings, 1):
prompt += f"{i}. {m}\n"
if wordnet_data.get("synonyms"):
prompt += f"同义词：{', '.join(wordnet_data['synonyms'][:8])}\n"
prompt += (
f"\n硬性要求：「中文意思」必须恰好输出 {expected_count} 条，"
f"与上面释义一一对应，编号 1 到 {expected_count} 都要有，不得合并、不得省略。\n"
)
else:
prompt += (
f"（WordNet 未找到该单词，请根据你自己的知识作答，"
f"「中文意思」至少输出 {expected_count} 条）\n"
)
example_lines = "\n".join(
f"- n. 中文义项{i}" for i in range(1, expected_count + 1)
)
prompt += (
"\n请严格按以下格式输出，不要添加其他标题或说明：\n"
"**中文意思**\n"
f"{example_lines}\n"
"**用法**\n"
"- 固定搭配：give a {word} 做{word} / 发表{word}\n"
"- 常见用法：make a {word} to... 向...做{word}\n"
"**其他形式**\n"
"- 单词（词性. 中文含义）\n"
"**例句**\n"
"- 英文例句（中文翻译）\n"
).format(word=word)
url = f"{OLLAMA_BASE}/api/generate"
payload = {
"model": OLLAMA_MODEL,
"prompt": prompt,
"stream": False,
"options": {
"temperature": 0.3,
"num_predict": 2048,
},
}
resp = requests.post(url, json=payload, timeout=180)
resp.raise_for_status()
data = resp.json()
print("done_reason:", data.get("done_reason"))
print("eval_count:", data.get("eval_count"))
return data["response"]
def parse_ollama_output(text: str) -> dict:
"""把模型输出解析成结构化字典。"""
result = {
"meaning": [],
"collocations": [],
"usage": [],
"forms": [],
"examples": [],
}
parts = re.split(r"\*\*(.+?)\*\*", text)
current_title = None
for i, part in enumerate(parts):
if i == 0:
continue
if i % 2 == 1:
current_title = part.strip()
else:
body = part.strip()
if not current_title or not body:
continue
if "中文意思" in current_title or "意思" in current_title:
result["meaning"] = [
line.strip().lstrip("-").strip()
for line in body.splitlines()
if line.strip()
]
elif current_title == "用法":
for line in body.splitlines():
line = line.strip()
if not line:
continue
clean = line.lstrip("-").strip()
if not clean:
continue
if clean.startswith("固定搭配：") or clean.startswith("固定搭配:"):
result["collocations"].append(clean)
else:
result["usage"].append(clean)
elif "其他形式" in current_title or "形式" in current_title:
result["forms"] = [
line.strip().lstrip("-").strip()
for line in body.splitlines()
if line.strip()
]
elif "例句" in current_title:
result["examples"] = [
line.strip().lstrip("-").strip()
for line in body.splitlines()
if line.strip()
]
return result
def build_markdown_block(word: str, wn_data: dict, ollama_text: str) -> str:
"""组装最终的 HTML + Markdown 块。"""
ollama_parsed = parse_ollama_output(ollama_text)
if ollama_parsed and ollama_parsed["meaning"]:
meanings_block = "\n".join(f"- {m}" for m in ollama_parsed["meaning"])
elif wn_data and wn_data.get("meanings"):
meanings_block = "\n".join(f"- {m}" for m in wn_data["meanings"])
else:
meanings_block = "- （未获取到释义，请手动补充）"
usage_lines = []
if ollama_parsed and ollama_parsed["collocations"]:
usage_lines.extend(ollama_parsed["collocations"])
if ollama_parsed and ollama_parsed["usage"]:
usage_lines.extend(ollama_parsed["usage"])
if not usage_lines:
usage_lines.append("（待补充）")
usage_block = "\n".join(f"- {u}" for u in usage_lines)
forms_lines = []
if ollama_parsed and ollama_parsed["forms"]:
forms_lines.extend(ollama_parsed["forms"])
if wn_data and wn_data.get("antonyms"):
forms_lines.append(f"反义词：{', '.join(wn_data['antonyms'][:8])}")
if not forms_lines:
forms_lines.append("（待补充）")
forms_block = "\n".join(f"- {f}" for f in forms_lines)
examples = []
if ollama_parsed and ollama_parsed["examples"]:
examples.extend(ollama_parsed["examples"][:3])
elif wn_data and wn_data.get("examples"):
examples.extend(wn_data["examples"][:3])
if not examples:
examples.append("（待补充）")
speak_buttons = ""
for ex in examples:
clean = re.sub(r"[（(].*?[）)]", "", ex).strip()
if clean:
speak_buttons += (
f'- <button class="speak-btn" data-text="{clean}">{clean}</button>'
)
if "（" in ex:
speak_buttons += ex[ex.rfind("（"):]
elif "(" in ex:
speak_buttons += ex[ex.rfind("("):]
speak_buttons += "\n"
else:
speak_buttons += f"- {ex}\n"
block = f"""<div class="word-item" data-word="{word}">
### {word}
**意思**
{meanings_block}
**用法**
{usage_block}
**其他形式**
{forms_block}
**例句**
{speak_buttons.rstrip()}
</div>
"""
return block
def main():
if len(sys.argv) < 2:
print("用法：python3 add-word.py <单词>")
sys.exit(1)
word = sys.argv[1].strip().lower()
print(f"正在查询 WordNet：{word}")
wn_data = lookup_wordnet(word)
print(f"WordNet 义项数：{len(wn_data.get('meanings', []))}")
print(f"正在调用 Ollama（模型：{OLLAMA_MODEL}）...")
ollama_text = ask_ollama(word, wn_data)
block = build_markdown_block(word, wn_data, ollama_text)
with open(TARGET_FILE, "a", encoding="utf-8") as f:
f.write(block + "\n")
print(f"已追加到：{TARGET_FILE}")
if __name__ == "__main__":
main()
|
