backend: add the reference library (参考文献)

One table per document the writer has read, and one for its contributors:

    reference          id, type, title, subtitle, title_translated,
                       container_title, volume, issue, pages, publisher,
                       publisher_place, event_*, publication_year/month/day,
                       accessed_date, doi, isbn, issn, url, patent_*,
                       standard_number, report_number, arxiv_id, abstract,
                       keywords, note, citation_key, language
    reference_author   reference_id, sort, name, family_name, given_name, role

The field set is the intersection of the formats a paper has to survive, taken
from the CSL JSON schema (103 fields, 45 types — what Zotero, Mendeley and
Pandoc exchange) against GB/T 7714-2015, BibTeX and RIS. The mapping is on
`app.models.reference` rather than in a commit message, because the next person
to add a column needs it more than this one does. Three decisions worth naming:

* **Partial dates.** Year, month and day are separate nullable integers, which
  is CSL's `issued`: a book has a year, a newspaper article a day, a patent its
  公告日期, and nothing invents the rest.
* **Contributors are rows, not a column.** A citation prints an ordered list
  with 编 / 译 marks, and styles rewrite names (「张三」 / "Zhang, S." /
  "S. Zhang"), so the printed form and its two halves are stored separately —
  display reads `name`, a reformatting style reads the halves, and an
  institutional author has neither.
* **`volume` / `issue` / `pages` are strings.** Journals print `12(Suppl.)`,
  `S1` and `e12345`, and a citation reproduces what is printed.

The library is global: the same article is cited from several papers, so the
entry cannot belong to one of them. The citation stays on the paper side in
`paper_sentence_reference`, which now carries a real foreign key to `reference`
(`ON DELETE RESTRICT`) — the column had been a loose integer waiting for this
table. A sentence may quote several documents and a document may be quoted by
many sentences, which is why the relation is a table and not a column on
`paper_sentence`.

Refusals are raised once, in the data layer, as `UnknownReference`, and
translated where each caller speaks: `400` with the offending ids for REST (one
exception handler, no route carries the check), a readable tool error for MCP
(one wrapper in the tool registry). Neither entry point can skip it, so the
foreign key never has to be the thing that complains.

API: `/references` CRUD plus batch delete, and `/references/{id}/citations` for
the reverse lookup — "where have I used this article?" — which is what tells a
writer whether an entry is safe to delete.

Verified: 29 new checks in `scripts/smoke_references.py` (partial dates, DOI
normalisation, author order and roles, search by author/type/year/cited, the
citation round trip, 400 on a phantom id, 409 on a cited entry), the paper
suite's 45 and the MCP suite's 47 still pass. The MCP suite now cites a real
entry instead of a phantom id, which is what an agent has to do now.
This commit is contained in:
2026-09-22 17:13:32 +08:00
parent 103b5544ba
commit d7bb777264
18 changed files with 1966 additions and 16 deletions
+260
View File
@@ -0,0 +1,260 @@
"""End-to-end check of the reference library, against a running API.
Run it while the backend is up::
.venv/bin/python scripts/smoke_references.py
It walks the whole loop the library exists for — add literature, find it again
by author and by type, link a sentence to it, look the citation up from the
reference side, be refused when deleting something in use, then clean up — and
exits non-zero on the first failed expectation. Everything it creates is
deleted, including the throwaway paper it cites from.
"""
from __future__ import annotations
import json
import sys
import urllib.error
import urllib.parse
import urllib.request
from typing import Any
BASE = "http://127.0.0.1:8000/api"
_checks = 0
def call(method: str, path: str, body: Any = None, expect: int = 200) -> Any:
"""Perform one API call and assert its status code."""
data = json.dumps(body).encode() if body is not None else None
url = BASE + urllib.parse.quote(path, safe="/?&=%")
request = urllib.request.Request(
url,
data=data,
method=method,
headers={"Content-Type": "application/json"},
)
try:
with urllib.request.urlopen(request, timeout=20) as response:
status = response.status
payload = response.read().decode()
except urllib.error.HTTPError as error:
status = error.code
payload = error.read().decode()
if status != expect:
raise AssertionError(f"{method} {path} -> {status}, expected {expect}: {payload}")
return json.loads(payload) if payload else None
def check(label: str, condition: bool, detail: str = "") -> None:
"""Assert one expectation and report it."""
global _checks
_checks += 1
if not condition:
raise AssertionError(f"{label}: {detail}")
print(f" ok {label}")
def main() -> int:
created_references: list[int] = []
created_papers: list[int] = []
try:
print("1. add an article with four authors")
article = call(
"POST",
"/references",
{
"type": "journal-article",
"title": "技术惯性与冲量的耦合机制",
"subtitle": "基于科学与技术融合网络的实证分析",
"title_translated": "Coupling inertia and impulse in technology evolution",
"container_title": "Technological Forecasting and Social Change",
"container_title_short": "Technol. Forecast. Soc. Change",
"volume": "198",
"issue": "3",
"pages": "122-141",
"publication_year": 2024,
"publication_month": 5,
"doi": "https://doi.org/10.1016/j.techfore.2024.123456",
"language": "zh",
"keywords": "技术演化, 惯性, 冲量",
"authors": [
{"name": "张三", "family_name": "张", "given_name": "三"},
{"name": "李四"},
{"name": "Wang Wu", "family_name": "Wang", "given_name": "Wu"},
{"name": "世界卫生组织", "role": "compiler"},
],
},
expect=201,
)
created_references.append(article["id"])
check("the DOI lost its URL prefix", article["doi"] == "10.1016/j.techfore.2024.123456", article["doi"])
check("four contributors stored", len(article["authors"]) == 4)
check(
"they keep the order they were sent in",
[a["name"] for a in article["authors"]] == ["张三", "李四", "Wang Wu", "世界卫生组织"],
str([a["name"] for a in article["authors"]]),
)
check(
"sort is numbered from 1",
[a["sort"] for a in article["authors"]] == [1, 2, 3, 4],
str([a["sort"] for a in article["authors"]]),
)
check("roles survive", article["authors"][3]["role"] == "compiler")
check("partial date kept as sent", article["publication_year"] == 2024 and article["publication_month"] == 5)
check("no day invented", article["publication_day"] is None)
check("not cited yet", article["citation_count"] == 0)
print("2. add a patent, which needs a country and a number")
patent = call(
"POST",
"/references",
{
"type": "patent",
"title": "一种基于多层聚类的人工智能技术识别方法",
"patent_country": "中国",
"patent_number": "CN114000000A",
"publication_year": 2023,
"authors": [{"name": "张三"}],
},
expect=201,
)
created_references.append(patent["id"])
check("patent stored", patent["type"] == "patent" and patent["patent_number"] == "CN114000000A")
print("3. a webpage without a URL is refused")
call(
"POST",
"/references",
{"type": "web-page", "title": "某网页"},
expect=422,
)
check("web page without URL refused", True)
print("4. a bad type, a phantom citation target and a missing title are refused")
call("POST", "/references", {"type": "blog-post", "title": "x"}, expect=422)
call("POST", "/references", {"title": " "}, expect=422)
call(
"POST",
"/references",
{"title": "引用了不存在的文献", "authors": [{"name": "甲"}], "type": "book"},
expect=201,
)
probe = call("GET", "/references?keyword=引用了不存在")
created_references.append(probe["items"][0]["id"])
check("a book needs nothing else", probe["total"] == 1)
print("5. find it again: by author, by type, by year")
by_author = call("GET", "/references?keyword=Wang")
check("search matches an author name", by_author["total"] == 1, str(by_author["total"]))
check(
"the row prints an author line",
by_author["items"][0]["authors_line"].startswith("张三, 李四, Wang Wu"),
by_author["items"][0]["authors_line"],
)
by_type = call("GET", "/references?type=patent")
check("filter by type", by_type["total"] == 1, str(by_type["total"]))
by_year = call("GET", "/references?year=2024")
check("filter by year", by_year["total"] == 1, str(by_year["total"]))
uncited = call("GET", "/references?cited=false")
check("filter by not-yet-cited", uncited["total"] == len(created_references))
print("6. edit it: replace the contributor list, change a field")
updated = call(
"PATCH",
f"/references/{article['id']}",
{
"pages": "122-145",
"authors": [
{"name": "张三", "family_name": "张", "given_name": "三"},
{"name": "李四"},
],
},
)
check("pages updated", updated["pages"] == "122-145")
check("author list replaced wholesale", len(updated["authors"]) == 2)
check("untouched fields kept", updated["container_title"] == "Technological Forecasting and Social Change")
print("7. cite it from a sentence")
templates = call("GET", "/templates?page_size=1")["items"]
paper = call(
"POST",
"/papers",
{"title": "参考文献冒烟测试论文", "template_id": templates[0]["id"]},
expect=201,
)
created_papers.append(paper["id"])
outline = call("GET", f"/templates/{templates[0]['id']}")
position = outline["placements"][0]["sort"]
call(
"PUT",
f"/papers/{paper['id']}/paragraphs/{position}",
{
"sentences": [
{
"content": "技术惯性约束了同时具有高冲量的可能。",
"citations": [
{"reference_id": article["id"], "quote": "高惯性技术几乎不可能同时具有高冲量。"}
],
}
]
},
)
check("a citation to a real reference is accepted", True)
missing = call(
"PUT",
f"/papers/{paper['id']}/paragraphs/{position}",
{"sentences": [{"content": "x", "citations": [{"reference_id": 999999, "quote": "y"}]}]},
expect=400,
)
check("a citation to nothing is refused by name", "999999" in missing["detail"], missing["detail"])
print("8. the reverse lookup: which sentences use this article?")
citations = call("GET", f"/references/{article['id']}/citations")
check("one citation found", len(citations) == 1, str(len(citations)))
check("it names the sentence", "技术惯性约束" in citations[0]["sentence_content"])
check("and the paper", citations[0]["paper_id"] == paper["id"])
check("and the paragraph heading", citations[0]["paragraph_name"] is not None, str(citations[0]))
cited = call("GET", "/references?cited=true")
check("the cited filter now finds it", cited["total"] == 1, str(cited["total"]))
print("9. a cited reference cannot be deleted; an uncited one can")
conflict = call("DELETE", f"/references/{article['id']}", expect=409)
check("refusal explains itself", "正被论文引用" in conflict["detail"], conflict["detail"])
call("DELETE", f"/references/{patent['id']}", expect=204)
created_references.remove(patent["id"])
check("an uncited reference deletes cleanly", True)
print("10. clean up")
call("DELETE", f"/papers/{paper['id']}", expect=204)
created_papers.clear()
for reference_id in created_references:
call("DELETE", f"/references/{reference_id}", expect=204)
created_references.clear()
remaining = call("GET", "/references")
check("nothing of ours is left", remaining["total"] == 0, str(remaining["total"]))
print(f"\nall {_checks} checks passed")
return 0
finally:
for paper_id in created_papers:
try:
call("DELETE", f"/papers/{paper_id}", expect=204)
except Exception as error: # noqa: BLE001 - cleanup must not mask the failure
print(f"cleanup of paper {paper_id} failed: {error}", file=sys.stderr)
for reference_id in created_references:
try:
call("DELETE", f"/references/{reference_id}", expect=204)
print(f"cleaned up reference {reference_id}")
except Exception as error: # noqa: BLE001
print(f"cleanup of reference {reference_id} failed: {error}", file=sys.stderr)
if __name__ == "__main__":
sys.exit(main())