利用DeepSeek和qwen 3.5辅助生成SQL优化方法幻灯片视频

一开始想用qwen 3.5来做，结果做成的是图形格式的PDF, 而且不能输出生成幻灯片的脚本，不太满意，于是让DeepSeek再做一版。
提示如下, 提纲是qwen输出的。

生成SQL优化方法介绍幻灯片。
优化维度：1. 索引优化基础与使用原则（合理创建B+树索引｜避免在索引列上使用函数或计算）；2. 查询语句写法规范与技巧（优先使用WHERE而非HAVING｜避免SELECT *｜善用LIMIT分页）；3. 表结构设计基础规范（遵循三范式但适度反范式｜合理选择数据类型｜主键必设且用INT/UUID）；4. 使用EXPLAIN分析执行计划（关注type、rows、key、Extra字段｜识别ALL/Using filesort等警告）；5. 避免全表扫描与常见性能陷阱（NULL值导致索引失效｜LIKE '%abc’无法走索引｜JOIN字段未建索引），最终凝练为“少查、快连、精索、常析、避坑”五大实战原则。
先输出第一个专题：索引。数据库中B+树索引的原理、创建、失效规避与实践验证方法；涵盖合理构建复合索引（如idx_user_status_time）、严防索引失效场景（禁止函数包裹、LIKE ‘%xxx’、隐式转换等）、通过EXPLAIN ANALYZE监控验证效果三大策略；深入解析索引底层结构、写法合规性对执行计划的影响，突出“索引要建得准、用得对、管得勤”。
用postgresql数据库举例，输出生成幻灯片的python代码。

首先根据代码中的提示安装

pip install manim Installing collected packages: svgelements, pyglm, pydub, glcontext, watchdog, srt, soupsieve, skia-pathops, screeninfo, pyglet, pycairo, networkx, moderngl, mapbox-earcut, manimpango, isosurfaces, decorator, av, audioop-lts, moderngl-window, cloup, beautifulsoup4, manim Successfully installed audioop-lts-0.2.2 av-16.1.0 beautifulsoup4-4.14.3 cloup-3.0.8 decorator-5.2.1 glcontext-3.0.0 isosurfaces-0.1.2 manim-0.19.2 manimpango-0.6.1 mapbox-earcut-2.0.0 moderngl-5.12.0 moderngl-window-3.1.1 networkx-3.6.1 pycairo-1.29.0 pydub-0.25.1 pyglet-2.1.13 pyglm-2.8.3 screeninfo-0.8.1 skia-pathops-0.9.2 soupsieve-2.8.3 srt-3.5.3 svgelements-1.9.6 watchdog-6.0.0 pip install manim-slides Installing collected packages: XlsxWriter, rtoml, qtpy, MarkupSafe, python-pptx, jinja2, click-default-group, pydantic-extra-types, manim-slides Successfully installed MarkupSafe-3.0.3 XlsxWriter-3.2.9 click-default-group-1.2.4 jinja2-3.1.6 manim-slides-5.5.4 pydantic-extra-types-2.11.0 python-pptx-1.0.2 qtpy-2.4.3 rtoml-0.13.0

执行命令行, 他生成的脚本有错，

manim -pql index_slides.py IndexOptimization C:\Users\lt\AppData\Local\Programs\Python\Python313\Lib\site-packages\pydub\utils.py:170: RuntimeWarning: Couldn't find ffmpeg or avconv - defaulting to ffmpeg, but may not work warn("Couldn't find ffmpeg or avconv - defaulting to ffmpeg, but may not work", RuntimeWarning) ... TypeError: Code.__init__() got an unexpected keyword argument 'code' TypeError: Code.__init__() got an unexpected keyword argument 'insert_line_no'

把错误信息提供给他，他把code改成text后仍然不对，自己到manim 官方文档页面查询，得知了正确的参数名是code_string，同样的，insert_line_no应该是add_line_numbers，并且不存在font_size参数，都改正后的脚本如下：

frommanimimport*frommanim_slidesimportSlide# 配置中文字体，请根据系统环境调整config.tex_template.add_to_preamble(r"\usepackage{ctex}")config.tex_template.add_to_preamble(r"\usepackage{xcolor}")classIndexOptimization(Slide):"""PostgreSQL 索引优化专题幻灯片 - 索引篇"""defconstruct(self):# ---------- 标题页 ----------title=Text("PostgreSQL 索引优化实战",font_size=48,color=BLUE)subtitle=Text("第一专题：索引",font_size=36,color=GRAY)authors=Text("少查·快连·精索·常析·避坑",font_size=28,color=GREEN)VGroup(title,subtitle,authors).arrange(DOWN,buff=0.5)self.play(Write(title))self.play(FadeIn(subtitle,shift=UP))self.play(FadeIn(authors,shift=UP))self.wait(1)self.next_slide()# 进入下一页# 清除当前画面self.clear()# ---------- 1. B+树索引原理 ----------principle_title=Text("1. B+树索引原理",font_size=40,color=YELLOW).to_edge(UP)self.play(Write(principle_title))# 绘制简单的B+树结构root=Circle(radius=0.3,color=BLUE).shift(UP*3)root_label=Text("根页",font_size=20).move_to(root)leaf1=Circle(radius=0.3,color=GREEN).shift(LEFT*3)leaf2=Circle(radius=0.3,color=GREEN).shift(LEFT*1)leaf3=Circle(radius=0.3,color=GREEN).shift(RIGHT*1)leaf4=Circle(radius=0.3,color=GREEN).shift(RIGHT*3)leaf_label1=Text("数据页1",font_size=16).next_to(leaf1,DOWN)leaf_label2=Text("数据页2",font_size=16).next_to(leaf2,DOWN)leaf_label3=Text("数据页3",font_size=16).next_to(leaf3,DOWN)leaf_label4=Text("数据页4",font_size=16).next_to(leaf4,DOWN)arrows=VGroup(Arrow(root.get_bottom(),leaf1.get_top(),buff=0.1),Arrow(root.get_bottom(),leaf2.get_top(),buff=0.1),Arrow(root.get_bottom(),leaf3.get_top(),buff=0.1),Arrow(root.get_bottom(),leaf4.get_top(),buff=0.1),)tree=VGroup(root,root_label,leaf1,leaf2,leaf3,leaf4,leaf_label1,leaf_label2,leaf_label3,leaf_label4,arrows)tree.scale(0.8).shift(DOWN*1.5)self.play(GrowFromCenter(root),Write(root_label))self.play(*[GrowFromCenter(leaf)forleafin[leaf1,leaf2,leaf3,leaf4]])self.play(*[Write(lab)forlabin[leaf_label1,leaf_label2,leaf_label3,leaf_label4]])self.play(*[GrowArrow(arr)forarrinarrows])# 添加解释文字explain=Text("B+树特点:\n- 所有数据在叶子节点\n- 叶子间双向链表连接\n- 高度平衡，3~4层即可支撑百万记录",font_size=24,color=GRAY,line_spacing=1.2).to_edge(RIGHT).shift(UP)self.play(Write(explain))self.wait(2)self.next_slide()# ---------- 2. 索引创建规范 ----------self.clear()create_title=Text("2. 索引创建规范",font_size=40,color=YELLOW).to_edge(UP)self.play(Write(create_title))# 表结构示例table_code=Code(code_string=''' CREATE TABLE users ( id BIGSERIAL PRIMARY KEY, status INT NOT NULL, created_at TIMESTAMP NOT NULL, email VARCHAR(255) ); ''',language="sql",#font_size=24,background="window",add_line_numbers=False).shift(LEFT*3).scale(0.8)# 复合索引推荐idx_code=Code(code_string=''' -- 推荐复合索引 CREATE INDEX idx_user_status_time ON users(status, created_at); -- 适合查询: SELECT * FROM users WHERE status = 1 ORDER BY created_at DESC; ''',language="sql",#font_size=24,background="window",add_line_numbers=False).next_to(table_code,DOWN,buff=0.5)self.play(FadeIn(table_code,shift=UP))self.play(FadeIn(idx_code,shift=UP))# 附加说明note=Text("原则:\n- 等值条件列放前面\n- 排序/范围列放后面\n- 索引列不宜过多(5列内)",font_size=24,color=BLUE,line_spacing=1.5).to_edge(RIGHT)self.play(Write(note))self.wait(2)self.next_slide()# ---------- 3. 索引失效场景 ----------self.clear()invalid_title=Text("3. 严防索引失效场景",font_size=40,color=YELLOW).to_edge(UP)self.play(Write(invalid_title))# 用表格展示失效场景table=Table([["函数包裹","WHERE func(col) = value"],["LIKE '%xxx'","WHERE name LIKE '%abc'"],["隐式转换","WHERE phone = 123456 (phone是varchar)"],["复合索引跳跃列","跳过前导列直接查created_at"]],col_labels=[Text("失效场景"),Text("错误示例")],include_outer_lines=True,line_config={"stroke_width":1,"color":GRAY},element_to_mobject_config={"font_size":20}).scale(0.7).shift(UP*1)self.play(Create(table))# 添加红色警示warning=Text("后果: 全表扫描(ALL) | 性能急剧下降",font_size=28,color=RED).shift(DOWN*2)self.play(Write(warning))self.wait(2)self.next_slide()# ---------- 4. 通过EXPLAIN ANALYZE验证 ----------self.clear()explain_title=Text("4. EXPLAIN ANALYZE 验证效果",font_size=40,color=YELLOW).to_edge(UP)self.play(Write(explain_title))# 展示执行计划对比plan_good=Code(code_string='''-- 正确使用索引的计划 => EXPLAIN ANALYZE SELECT * FROM users WHERE status = 1 ORDER BY created_at; -------------------------------------------------------------------- Index Scan using idx_user_status_time on users (cost=0.15..8.17 rows=1 width=8) (actual time=0.024..0.025 rows=1 loops=1) Planning Time: 0.123 ms Execution Time: 0.045 ms''',language="sql",#font_size=20,background="window",add_line_numbers=False).shift(UP*0.5).scale(0.7)plan_bad=Code(code_string='''-- 函数包裹导致索引失效 => EXPLAIN ANALYZE SELECT * FROM users WHERE EXTRACT(YEAR FROM created_at) = 2023; -------------------------------------------------------------------- Seq Scan on users (cost=0.00..22.75 rows=4 width=8) (actual time=0.078..0.139 rows=4 loops=1) Planning Time: 0.234 ms Execution Time: 0.176 ms''',language="sql",#font_size=20,background="window",add_line_numbers=False).next_to(plan_good,DOWN,buff=0.3).scale(0.7)self.play(FadeIn(plan_good,shift=LEFT))self.play(FadeIn(plan_bad,shift=RIGHT))# 高亮关键指标circle_good=Circle(color=GREEN,radius=0.5).move_to(plan_good.get_center()+DOWN*0.5)circle_bad=Circle(color=RED,radius=0.5).move_to(plan_bad.get_center()+DOWN*0.5)self.play(Create(circle_good),Create(circle_bad))note=Text("关注: type(Index Scan vs Seq Scan), actual time, rows",font_size=24).to_edge(DOWN)self.play(Write(note))self.wait(2)self.next_slide()# ---------- 5. 总结与实战原则 ----------self.clear()summary_title=Text("索引优化五大原则",font_size=44,color=YELLOW).to_edge(UP)self.play(Write(summary_title))principles=VGroup(Text("1️⃣ 建得准: 根据查询模式设计复合索引 (如idx_user_status_time)",font_size=28),Text("2️⃣ 用得对: 避免函数、隐式转换、前导模糊",font_size=28),Text("3️⃣ 管得勤: 定期用EXPLAIN ANALYZE验证, 删除冗余索引",font_size=28),Text("4️⃣ 覆盖索引: 考虑包含列(INCLUDE)减少回表",font_size=28),Text("5️⃣ 监控: pg_stat_user_indexes 查看使用率",font_size=28),).arrange(DOWN,aligned_edge=LEFT,buff=0.3).shift(UP*0.5)forpinprinciples:self.play(Write(p,lag_ratio=0.1))self.wait(0.3)footer=Text("下一专题预告: 查询语句写法规范与技巧",font_size=24,color=GRAY).to_edge(DOWN)self.play(Write(footer))self.wait(3)# 如果要生成视频幻灯片，需安装 manim 和 manim-slides，然后运行:# manim -pql index_slides.py IndexOptimization# manim-slides indexoptimization

另外，下载ffmpeg, 将压缩包解压到C:\d\ffmpeg\bin，添加到PATH环境变量。

wget https://www.gyan.dev/ffmpeg/builds/ffmpeg-release-essentials.7z Saving to: 'ffmpeg-release-essentials.7z' set path=%path%;C:\d\ffmpeg\bin

再次执行就成功生成了mp4格式的幻灯片。

INFO Rendered IndexOptimization scene.py:278 Played 39 animations [02/17/26 17:19:55] INFO Previewed File at: file_ops.py:237 'C:\d\media\videos\index_slides\480p15\IndexOptimization.mp4' [02/17/26 17:19:58] INFO Generated 6 slides to 'C:\d\slides\files\IndexOptimization' base.py:599 [02/17/26 17:19:58] INFO Generated 6 slides to 'C:\d\slides\files\IndexOptimization' base.py:599 INFO Slide 'IndexOptimization' configuration written in base.py:611 'C:\d\slides\IndexOptimization.json' INFO Slide 'IndexOptimization' configuration written in base.py:611 'C:\d\slides\IndexOptimization.json'

第二个命令行还是报错，看提示是缺少qt包，暂时不管了。

manim-slides indexoptimization Traceback (most recent call last): ... qtpy.QtBindingsNotFoundError: No Qt bindings could be found