CheshireMew/100x-learning / archive/writing-capability-retired-2026-09-02
100x-learning
Verified帮助用户读懂材料、研究主题、解释概念、把知识用于真实问题,并直接写成文章;也能把字幕和自动转写稿整理成忠实、可读的文章式全文并写入私人知识库,在用户明确要求时再讲解材料或提炼知识,接入视频和社交来源,筛选分享内容,根据发布反馈继续选题和改进内容,以及初始化、检查和维护本机私人知识库、内容案例、钩子与写作记忆。适用于字幕整稿、划重点、系统学习、分轮处理长材料、内容审查、文章写作和知识沉淀;不负责短内容生成、媒体制作、普通翻译、广告投放、销售页、邮件营销、品牌全案、完整营销策略和实际发布。
tests/test_writing_memory.py
View on GitHubfrom __future__ import annotations
import json
import subprocess
import sys
import tempfile
import unittest
from pathlib import Path
from scripts.content_case_library import add_case, load_library as load_cases
from scripts.private_library import initialize_library
from scripts.writing_memory import (
CONFIG_RELATIVE,
INDEX_RELATIVE,
MemoryError,
discover_records,
index_is_current,
load_index,
search_memory,
write_index,
)
SCRIPT_PATH = Path(__file__).resolve().parents[1] / "scripts" / "writing_memory.py"
def _write(path: Path, text: str) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(text, encoding="utf-8")
class WritingMemoryTests(unittest.TestCase):
def setUp(self) -> None:
self.temp_dir = tempfile.TemporaryDirectory()
self.temp_root = Path(self.temp_dir.name)
self.layout, _, _ = initialize_library(
self.temp_root / "private-library",
self.temp_root / "config.json",
)
self.project_root = self.layout.root
self.blog_root = self.project_root / "20-Sources" / "Articles" / "Published"
self.output_root = self.layout.writing_outputs
self.blog_root.mkdir(parents=True, exist_ok=True)
(self.project_root / CONFIG_RELATIVE).write_text(
json.dumps(
{
"published_article_roots": ["20-Sources/Articles/Published"],
},
ensure_ascii=False,
),
encoding="utf-8",
)
def tearDown(self) -> None:
self.temp_dir.cleanup()
def test_publication_history_keeps_ai_work_but_does_not_inherit_voice(self) -> None:
_write(
self.blog_root / "published.md",
"""---
authorship: "本人主导"
content_type: "教程与操作指南"
source_url: "https://example.com/published/"
---
# 已发布版本
旧正文。
""",
)
_write(
self.blog_root / "ai-draft.md",
"""---
authorship: "AI 主笔"
content_type: "教程与操作指南"
source_url: "https://example.com/ai/"
---
# AI 草稿
不能进入作者证据。
""",
)
_write(
self.output_root / "Articles" / "confirmed.md",
"""---
type: writing-output
status: final
source: published-article
author: 柴郡
updated: 2026-07-29
published_url: "https://example.com/published/"
---
# 用户确认版本
新正文,应该成为唯一活动记录。
""",
)
_write(
self.output_root / "Drafts" / "draft.md",
"""---
type: writing-output
status: draft
source: user-confirmed
format: original
---
# 未完成草稿
不能进入作者证据。
""",
)
records, receipt = discover_records(self.project_root)
self.assertEqual(2, len(records))
by_url = {record.source_url: record for record in records}
confirmed = by_url["https://example.com/published/"]
ai_work = by_url["https://example.com/ai/"]
self.assertEqual("用户确认版本", confirmed.title)
self.assertEqual("教程与操作指南", confirmed.content_type)
self.assertEqual("writing-output", confirmed.source_kind)
self.assertEqual("unknown", confirmed.writing_origin)
self.assertFalse(confirmed.voice_eligible)
self.assertEqual("ai-generated", ai_work.writing_origin)
self.assertFalse(ai_work.voice_eligible)
self.assertEqual(2, receipt.accepted_blog)
self.assertEqual(1, receipt.merged_by_url)
def test_formal_article_body_is_consumed_as_authored_evidence(self) -> None:
_write(
self.blog_root / "case.md",
"""---
authorship: "本人主导"
content_type: "个人观察与实测"
source_url: "https://example.com/case/"
---
# 本人文章
这是实际发布正文。
""",
)
records, _ = discover_records(self.project_root)
write_index(self.project_root, records)
hits = search_memory(
library_root=self.project_root,
records=load_index(self.project_root),
purpose="novelty",
query="实际发布正文",
format_name="article",
content_type=None,
limit=1,
)
self.assertIn("实际发布正文", hits[0].opening)
def test_explicit_writing_metadata_adds_a_case_without_account_or_source(self) -> None:
case_inputs = self.temp_root / "case-inputs"
eligible_input = case_inputs / "eligible.md"
unknown_input = case_inputs / "unknown.md"
ordinary_input = case_inputs / "ordinary.md"
_write(eligible_input, "我真正想减少的是上下文切换,而不只是点击次数。")
_write(unknown_input, "这条内容明确保存为写作记录,但没有标注写法来源。")
_write(ordinary_input, "这只是普通参考案例,不进入写作记忆。")
existing: list = []
add_case(
self.layout,
existing,
kind="social",
input_path=eligible_input,
title="写作案例",
techniques=("亲历切入", "机制拆解"),
writing_format="product",
writing_purpose="个人观察与实测",
writing_origin="human-edited",
voice_eligible=True,
)
existing, issues = load_cases(self.layout)
self.assertFalse(issues)
add_case(
self.layout,
existing,
kind="social",
input_path=unknown_input,
title="写法未知",
techniques=("结果先行",),
writing_format="product",
writing_purpose="个人观察与实测",
)
existing, issues = load_cases(self.layout)
self.assertFalse(issues)
add_case(
self.layout,
existing,
kind="social",
input_path=ordinary_input,
title="普通案例",
techniques=("观点先行",),
)
records, receipt = discover_records(self.project_root)
self.assertEqual(2, len(records))
eligible = next(record for record in records if record.title == "写作案例")
unknown = next(record for record in records if record.title == "写法未知")
self.assertEqual("writing-case", eligible.source_kind)
self.assertEqual("product", eligible.format)
self.assertEqual("个人观察与实测", eligible.content_type)
self.assertEqual("", eligible.updated)
self.assertEqual("", eligible.source_url)
self.assertEqual("human-edited", eligible.writing_origin)
self.assertTrue(eligible.voice_eligible)
self.assertEqual("unknown", unknown.writing_origin)
self.assertFalse(unknown.voice_eligible)
self.assertEqual(3, receipt.scanned_cases)
self.assertEqual(2, receipt.accepted_cases)
write_index(self.project_root, records)
novelty_hits = search_memory(
library_root=self.project_root,
records=load_index(self.project_root),
purpose="novelty",
query="减少上下文切换",
format_name="product",
content_type=None,
limit=1,
)
self.assertIn("上下文切换", novelty_hits[0].opening)
self.assertNotIn("content-case-index", novelty_hits[0].ending)
voice_hits = search_memory(
library_root=self.project_root,
records=load_index(self.project_root),
purpose="voice",
query="",
format_name="product",
content_type="个人观察与实测",
limit=3,
)
self.assertEqual(["写作案例"], [hit.record.title for hit in voice_hits])
def test_marked_writing_case_requires_its_actual_writing_format(self) -> None:
case = (
self.project_root
/ "20-Sources"
/ "Social Posts"
/ "Content Cases"
/ "完整社交内容"
/ "缺少形态.md"
)
_write(
case,
"""# 缺少形态
## 原文全文
这是本人发布的产品介绍。
<!-- content-case-index
writing_origin: "human"
writing_techniques: ["结果先行"]
-->
""",
)
with self.assertRaisesRegex(MemoryError, "缺少 writing_format"):
discover_records(self.project_root)
def test_unknown_writing_memory_configuration_is_rejected(self) -> None:
_write(
self.project_root / CONFIG_RELATIVE,
json.dumps(
{
"unexpected_field": [],
"published_article_roots": ["20-Sources/Articles/Published"],
}
),
)
with self.assertRaisesRegex(ValueError, "不受支持的字段"):
discover_records(self.project_root)
def test_index_output_is_consumed_by_same_format_search(self) -> None:
_write(
self.output_root / "Articles" / "article.md",
"""---
type: writing-output
status: final
source: user-confirmed
author: 柴郡
updated: 2026-07-29
format: article
content_type: 个人观察与实测
---
# 文章
这个工具真正省下来的不是点击次数,而是来回切换上下文。
""",
)
_write(
self.output_root / "Social" / "X" / "original.md",
"""---
type: writing-output
status: final
source: user-confirmed
author: 柴郡
updated: 2026-07-28
format: original
content_type: 项目与产品介绍
---
# 独立帖
我试了一个新的写作工具,最后留下的是更完整的工作流。
""",
)
records, _ = discover_records(self.project_root)
index_path = write_index(self.project_root, records)
self.assertEqual(self.project_root / INDEX_RELATIVE, index_path)
self.assertTrue(index_is_current(self.project_root, records))
indexed = load_index(self.project_root)
hits = search_memory(
library_root=self.project_root,
records=indexed,
purpose="novelty",
query="工具如何减少上下文切换",
format_name="article",
content_type=None,
limit=3,
)
self.assertEqual(1, len(hits))
self.assertEqual("article", hits[0].record.format)
self.assertIn("上下文", hits[0].opening)
def test_cli_build_then_search_reads_generated_index(self) -> None:
_write(
self.output_root / "Social" / "X" / "thread.md",
"""---
type: writing-output
status: final
source: published-thread
author: 柴郡
updated: 2026-07-29
format: thread
published_url: "https://x.com/example/status/1"
---
# 一条旧 Thread
先看问题怎样发生,再看哪个机制真正改变了结果。
""",
)
build = subprocess.run(
[
sys.executable,
str(SCRIPT_PATH),
"--library-root",
str(self.project_root),
"build-index",
],
check=False,
capture_output=True,
text=True,
encoding="utf-8",
)
self.assertEqual(0, build.returncode, build.stderr)
receipt = json.loads(build.stdout)
self.assertEqual(1, receipt["records"])
search = subprocess.run(
[
sys.executable,
str(SCRIPT_PATH),
"--library-root",
str(self.project_root),
"search",
"--purpose",
"novelty",
"--query",
"问题发生后哪个机制改变结果",
"--format",
"thread",
],
check=False,
capture_output=True,
text=True,
encoding="utf-8",
)
self.assertEqual(0, search.returncode, search.stderr)
self.assertIn("数据来源:发布记录索引", search.stdout)
self.assertIn("一条旧 Thread", search.stdout)
def test_changed_source_makes_index_stale(self) -> None:
output = self.output_root / "Articles" / "article.md"
_write(
output,
"""---
type: writing-output
status: final
source: user-confirmed
author: 柴郡
---
# 原文
第一版。
""",
)
records, _ = discover_records(self.project_root)
write_index(self.project_root, records)
self.assertTrue(index_is_current(self.project_root, records))
_write(
output,
"""---
type: writing-output
status: final
source: user-confirmed
author: 柴郡
---
# 原文
第二版。
""",
)
current, _ = discover_records(self.project_root)
self.assertFalse(index_is_current(self.project_root, current))
def test_missing_same_format_is_a_normal_empty_result(self) -> None:
_write(
self.output_root / "Articles" / "article.md",
"""---
type: writing-output
status: final
source: user-confirmed
author: 柴郡
format: article
---
# 文章
只有文章证据。
""",
)
build = subprocess.run(
[
sys.executable,
str(SCRIPT_PATH),
"--library-root",
str(self.project_root),
"build-index",
],
check=False,
capture_output=True,
text=True,
encoding="utf-8",
)
self.assertEqual(0, build.returncode, build.stderr)
search = subprocess.run(
[
sys.executable,
str(SCRIPT_PATH),
"--library-root",
str(self.project_root),
"search",
"--purpose",
"voice",
"--format",
"thread",
],
check=False,
capture_output=True,
text=True,
encoding="utf-8",
)
self.assertEqual(0, search.returncode, search.stderr)
self.assertIn("作者声音候选(0 条)", search.stdout)
self.assertIn("不拿发布历史硬凑", search.stdout)
def test_voice_search_rejects_topic_queries(self) -> None:
_write(
self.output_root / "Social" / "X" / "voice.md",
"""---
type: writing-output
status: final
source: user-confirmed
format: product
writing_origin: human-edited
voice_eligible: true
---
# 可用声音
这是经过确认的表达。
""",
)
records, _ = discover_records(self.project_root)
with self.assertRaisesRegex(MemoryError, "主题词只用于 novelty"):
search_memory(
library_root=self.project_root,
records=records,
purpose="voice",
query="币安 Robinhood Chain",
format_name="product",
content_type=None,
limit=1,
)
def test_removed_writing_formats_are_rejected_by_the_cli(self) -> None:
for removed_format in ("reply", "newsletter"):
with self.subTest(removed_format=removed_format):
result = subprocess.run(
[
sys.executable,
str(SCRIPT_PATH),
"--library-root",
str(self.project_root),
"search",
"--purpose",
"novelty",
"--query",
"任意内容",
"--format",
removed_format,
],
check=False,
capture_output=True,
text=True,
encoding="utf-8",
)
self.assertEqual(2, result.returncode)
self.assertIn("invalid choice", result.stderr)
if __name__ == "__main__":
unittest.main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |