feat: add compare, trends routes, embedder service, and phase5 tests

2026-06-05 23:32:06 +08:00
parent 2cfd1a8a9f
commit ba9afa212c
17 changed files with 2122 additions and 27 deletions
@@ -0,0 +1,115 @@
+"""论文对比页路由 — 多篇论文结构化字段并排对比。"""
+
+from __future__ import annotations
+
+import logging
+
+from fastapi import APIRouter, Depends, HTTPException, Query, Request
+from fastapi.templating import Jinja2Templates
+from sqlalchemy.orm import Session, joinedload
+
+from app.database import get_db
+from app.models import Paper
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter()
+templates = Jinja2Templates(directory="app/templates")
+
+
+@router.get("/compare")
+def compare_page(
+    request: Request,
+    ids: str = Query(default="", description="逗号分隔的 arxiv_id，最多 5 篇"),
+    db: Session = Depends(get_db),
+):
+    """论文对比页面。GET /compare?ids=id1,id2,id3"""
+    if not ids:
+        return templates.TemplateResponse(
+            request,
+            "compare.html",
+            {
+                "page_title": "论文对比",
+                "papers": [],
+                "error": None,
+            },
+        )
+
+    arxiv_ids = [i.strip() for i in ids.split(",") if i.strip()]
+
+    # 最多 5 篇
+    if len(arxiv_ids) > 5:
+        arxiv_ids = arxiv_ids[:5]
+
+    if not arxiv_ids:
+        return templates.TemplateResponse(
+            request,
+            "compare.html",
+            {
+                "page_title": "论文对比",
+                "papers": [],
+                "error": "请提供有效的论文 ID",
+            },
+        )
+
+    papers = (
+        db.query(Paper)
+        .filter(Paper.arxiv_id.in_(arxiv_ids))
+        .options(
+            joinedload(Paper.authors),
+            joinedload(Paper.tags),
+            joinedload(Paper.summary),
+            joinedload(Paper.summary_status),
+        )
+        .all()
+    )
+
+    # 按请求顺序排列
+    id_order = {aid: idx for idx, aid in enumerate(arxiv_ids)}
+    papers.sort(key=lambda p: id_order.get(p.arxiv_id, 999))
+
+    # 构建对比数据
+    compare_fields = [
+        ("title_zh", "中文标题"),
+        ("title_en", "英文标题"),
+        ("one_line", "一句话摘要"),
+        ("difficulty", "难度"),
+        ("motivation_problem", "研究问题"),
+        ("motivation_goal", "研究目标"),
+        ("motivation_gap", "研究差距"),
+        ("method_overview", "方法概述"),
+        ("method_key_idea", "关键思路"),
+        ("method_novelty", "新颖性"),
+        ("results", "实验结果"),
+        ("limitations", "局限与改进"),
+    ]
+
+    rows = []
+    for field_key, field_label in compare_fields:
+        cells = []
+        for paper in papers:
+            if field_key in ("title_zh", "title_en"):
+                val = getattr(paper, field_key, None) or ""
+            elif paper.summary:
+                val = getattr(paper.summary, field_key, None) or ""
+                # JSON 字段直接展示
+                if field_key == "results" and not val:
+                    val = paper.summary.results_main_json or ""
+                if field_key == "limitations" and not val:
+                    val = paper.summary.limitations_json or ""
+            else:
+                val = ""
+            cells.append(val)
+        rows.append({"key": field_key, "label": field_label, "cells": cells})
+
+    return templates.TemplateResponse(
+        request,
+        "compare.html",
+        {
+            "page_title": "论文对比",
+            "papers": papers,
+            "rows": rows,
+            "ids_param": ids,
+            "error": None,
+        },
+    )