import polars as pl

df = pl.read_parquet("./history_aio.parquet")
pl.__version__

'0.20.4'

heatmap_dat = (
    # 挑出 2023 年
    df.filter(pl.col("year") == 2023)
    # 按月、小时段，做 count
    .group_by("month", "hour")
    .agg(
        pl.count().alias("n"),
    )
    # 透视拉宽、补列，完整 24 小时
    .pivot(index="month", columns="hour", values="n")
    .with_columns(
        pl.lit(None).alias("4"), pl.lit(None).alias("5"), pl.lit(None).alias("6")
    )
    .select(["month"] + [str(c) for c in range(1, 24)])
    # 融化还原成长表
    .melt(id_vars="month", variable_name="hour", value_name="n")
    # 变换类型并排序
    .with_columns(pl.col("hour").cast(pl.Int8))
    .sort("month", "hour")
)

heatmap_dat.head()

bump_dat = (
    # 挑出有效类别行
    df.filter(pl.col("cat_name").is_not_null())
    # 按年、类，count 视频唯一量
    .group_by("year", "cat_name").agg(pl.n_unique("title").alias("n_video"))
    # 按量大小，排名
    .with_columns(
        pl.col("n_video").rank("ordinal", descending=True).over("year").alias("ranking")
    )
    # 按年排序
    .sort("year")
)

bump_dat.head()

# from bokeh.io import output_notebook

# output_notebook()

fig1 = heatmap_dat.plot.heatmap(
    x="hour", y="month", C="n", cmap="reds", title="Heatmap"
).options(toolbar=None, default_tools=[])

fig1

fig2 = (
    bump_dat.plot.line(x="year", y="ranking", by="cat_name")
    * bump_dat.plot.scatter(x="year", y="ranking", by="cat_name")
).opts(
    title="Bump Chart",
    toolbar=None,
    default_tools=[],
    invert_yaxis=True,
    legend_opts=dict(border_line_width=0, label_text_font_size="6pt", label_height=2),
    padding=0.05,
    height=480,
)

fig2

AttributeError: unexpected attribute 'border_width' to Legend, similar attributes are border_line_width, border_line_dash or label_width

fig3 = (fig1.opts(height=480) + fig2).opts(toolbar=None)
fig3

import altair as alt

alt.__version__

'5.2.0'

fig4 = (
    # 吃数据
    alt.Chart(heatmap_dat.fill_null(0))
    # 画方块
    .mark_rect()
    # 微调配置
    .encode(
        # x 及其数据类型
        x="hour:O",
        # y 及其数据类型、排序方式
        y=alt.Y("month:O", sort="descending"),
        # color 填充规则
        color=alt.condition(
            "datum.n != 0", alt.Color("n:Q").scale(scheme="reds"), alt.value("white")
        ),
        # 悬停信息
        tooltip=[
            alt.Tooltip("month"),
            alt.Tooltip("hour"),
            alt.Tooltip("n:Q", title="# of videos"),
        ],
    )
    # 补充定义
    .properties(title="Heatmap", width=500, height=300)
)

fig5 = (
    alt.Chart(bump_dat)
    # 画线、标点
    .mark_line(point=True)
    .encode(
        x="year:O",
        y=alt.Y("ranking:Q", sort="descending"),
        # 填充且干预排序
        color=alt.Color(
            "cat_name",
            sort=[
                "Science & Technology",
                "People & Blogs",
                "Film & Animation",
                "Education",
                "Gaming",
                "Entertainment",
                "Howto & Style",
                "Music",
                "Travel & Events",
                "Sports",
                "News & Politics",
                "Pets & Animals",
                "Autos & Vehicles",
                "Comedy",
                "Nonprofits & Activism",
            ],
        ),
        tooltip=[
            alt.Tooltip("year:O"),
            alt.Tooltip("cat_name"),
            alt.Tooltip("ranking:Q"),
        ],
    )
    .properties(
        title="Bump Chart",
        width=500,
        height=300,
    )
)

# 超方便的多图堆叠
# (fig1 & fig2)
fig6 = (fig4 | fig5).configure_axis(grid=False).properties(title="Bye, hell subplots.")
fig6

MaxRowsError: The number of rows in your dataset is greater than the maximum allowed (5000).

Try enabling the VegaFusion data transformer which raises this limit by pre-evaluating data
transformations in Python.
    >> import altair as alt
    >> alt.data_transformers.enable("vegafusion")

Or, see https://altair-viz.github.io/user_guide/large_datasets.html for additional information
on how to plot large datasets.

import plotly

plotly.__version__

'5.18.0'

import plotly.graph_objects as go
from plotly import express as px
from plotly.subplots import make_subplots

fig7 = (
    # 方块热力图在 plotly express 中表述为 imshow
    px.imshow(
        # 为了迎合 imshow 的表达方式，我们需要先揉捏一下数据
        heatmap_dat.pivot(index="month", columns="hour", values="n")
        .to_pandas()
        .set_index("month"),
        # 数据呈现方式
        labels=dict(x="hour", y="month", color="n_videos"),
        color_continuous_scale="reds",
        # 图表配置
        template="plotly_white",
        title="Heatmap",
    ).update_layout(
        xaxis1=dict(showgrid=False),
        yaxis1=dict(showgrid=False),
    )
)

fig7

def plotly_bump(df):
    fig = go.Figure()

    # Bump Chart 可以
    # 为每一个类别
    for name, data in df.group_by(["cat_name"]):
        name = name[0]
        trace = go.Scatter(
            x=data["year"],
            y=data["ranking"],
            mode="lines+markers",
            name=name,
            line=dict(width=2),
            marker=dict(size=10),
            text=name,
            hoverinfo="text",
            legendrank=data["ranking"][-1],
        )
        fig.add_trace(trace)

    fig.update_layout(
        title="Bump Chart",
        xaxis=dict(title="Year"),
        yaxis=dict(title="Rank"),
        legend=dict(title="Category"),
        yaxis_autorange="reversed",
        template="plotly_white",
        height=500,
        width=800,
        xaxis1=dict(showgrid=False),
        yaxis1=dict(showgrid=False),
    )

    return fig


fig8 = plotly_bump(bump_dat)
fig8

def plotly_subplots():
    fig = make_subplots(rows=1, cols=2, subplot_titles=["Heatmap", "Bump Chart"])

    # Heatmap
    heatmap_trace = go.Heatmap(
        x=heatmap_dat["hour"],
        y=heatmap_dat["month"],
        z=heatmap_dat["n"],
        hoverongaps=False,
        opacity=0.7,
        colorscale="reds",
        colorbar=dict(x=0.45, y=0.5, len=1.1),
    )
    fig.add_trace(
        heatmap_trace,
        row=1,
        col=1,
    )

    # Bump Chart
    for name, data in bump_dat.group_by(["cat_name"]):
        name = name[0]
        trace = go.Scatter(
            x=data["year"],
            y=data["ranking"],
            mode="lines+markers",
            name=name,
            line=dict(width=2),
            marker=dict(size=10),
            text=name,
            hoverinfo="text",
            legendrank=data["ranking"][-1],
            legendgroup="2",
        )
        fig.add_trace(trace, row=1, col=2)

    # Update layout
    fig.update_layout(
        title_text="Heatmap and Bump Chart Side by Side",
        template="plotly_white",
        xaxis1=dict(showgrid=False),
        yaxis1=dict(showgrid=False),
        yaxis2_autorange="reversed",
        height=400,
        width=1300,
    )

    return fig


fig9 = plotly_subplots()
fig9

import bokeh

bokeh.__version__

'3.3.3'

from bokeh.layouts import row
from bokeh.models import ColumnDataSource, Legend, LegendItem
from bokeh.plotting import figure, show, save
from bokeh.transform import factor_cmap, linear_cmap
from bokeh.io import output_notebook

output_notebook()

# 需要 Pandas DataFrame
fig10_dat = heatmap_dat.to_pandas()

# 控制热力图细节
fig10 = figure(
    title="Heatmap",
    # 小时
    x_range=[str(x) for x in range(0, 24)],
    # 月份，多一份留白，让每个小格子等宽等长
    y_range=[str(x) for x in range(1, 14)],
    x_axis_location="above",
    tools="",
    toolbar_location=None,
    tooltips=[("hour", "@hour"), ("month", "@month"), ("n", "@n")],
    width=600,
    height=300,
)

# 去掉干扰
fig10.grid.grid_line_color = None
fig10.axis.axis_line_color = None
fig10.axis.major_tick_line_color = None
fig10.outline_line_color = None

# 映射数据
fig10.rect(
    x="hour",
    y="month",
    width=1,
    height=1,
    source=fig10_dat,
    fill_color=linear_cmap(
        "n", "Reds256", high=fig10_dat.n.min(), low=fig10_dat.n.max(), nan_color="white"
    ),
    line_color=None,
)

# 展示图表
show(fig10)

# 初始化图表配置
fig11 = figure(
    title="Slope Graph",
    x_range=(2015.9, 2023.1),
    y_range=(17.5, 0.5),
    x_axis_location="above",
    tools="",
    toolbar_location=None,
    width=600,
    height=300,
    tooltips=[("category", "@cat_name"), ("ranking", "@ranking")],
)

# 预先设定 17 个分类的颜色映射关系
cmap6 = factor_cmap(
    "cat_name",
    "Category20_17",
    sorted(bump_dat["cat_name"].unique()),
)

# 预先设定好图例排序
fig11_legend_indexes = [
    "Science & Technology",
    "People & Blogs",
    "Film & Animation",
    "Education",
    "Gaming",
    "Entertainment",
    "Howto & Style",
    "Music",
    "Travel & Events",
    "Sports",
    "News & Politics",
    "Pets & Animals",
    "Autos & Vehicles",
    "Comedy",
    "Nonprofits & Activism",
    "Movies",
    "Trailers",
]

fig11_legend_items = []

# 分别绘制折线图和散点图，并把图例信息传出来
for name, data in bump_dat.group_by(["cat_name"]):
    src = data.to_pandas()
    name = name[0]
    line = fig11.line(
        x="year",
        y="ranking",
        source=src,
        line_width=2,
        line_color=cmap6.transform.palette[cmap6.transform.factors.index(name)],
    )
    scatter = fig11.scatter(
        x="year",
        y="ranking",
        source=src,
        size=8,
        color=cmap6.transform.palette[cmap6.transform.factors.index(name)],
    )

    fig11_legend_items.append(LegendItem(label=name, renderers=[line, scatter]))

# 排除干扰
fig11.grid.grid_line_color = None
fig11.axis.axis_line_color = None
fig11.axis.major_tick_line_color = None
fig11.outline_line_color = None
fig11.xaxis.minor_tick_out = 0
fig11.xaxis.major_tick_in = 0
fig11.yaxis.minor_tick_out = 0
fig11.yaxis.major_tick_in = 0

# 微调图例
fig11_legend_items_sorted = sorted(
    fig11_legend_items, key=lambda x: fig11_legend_indexes.index(x.label.value)
)
fig11.add_layout(Legend(items=fig11_legend_items_sorted, click_policy="mute"), "right")

# 排除干扰
fig11.legend.label_text_font_size = "6pt"
fig11.legend.glyph_height = 1
fig11.legend.spacing = 5
fig11.legend.label_height = 5
fig11.legend.location = (0, 0)
fig11.legend.background_fill_alpha = 0.6
fig11.legend.border_line_alpha = 0

# 展示图表
show(fig11)

fig12 = row(fig10, fig11)
show(fig12)

cd /Library/Frameworks/R.framework/Resources/lib
ln -sf libRblas.vecLib.dylib libRblas.dylib

conda install cython pybind11
pip install --no-binary :all: --no-use-pep517 numpy

 pip_interop_enabled: true

channels: [conda-forge]
pip_interop_enabled: true
repodata_threads: 4
pkgs_dirs:
  - ~/.local/python

micromamba install numpy "libblas=*=*accelerate"

library(tidyverse)

big_mac <- read_csv("https://github.com/TheEconomist/big-mac-data/releases/download/2021-01/big-mac-full-index.csv") |> 
  janitor::clean_names()

# A tibble: 1,386 x 19
   date       iso_a3 currency_code name       local_price dollar_ex
   <date>     <chr>  <chr>         <chr>            <dbl>     <dbl>
 1 2000-04-01 ARG    ARS           Argentina         2.5       1   
 2 2000-04-01 AUS    AUD           Australia         2.59      1.68
 3 2000-04-01 BRA    BRL           Brazil            2.95      1.79
 4 2000-04-01 CAN    CAD           Canada            2.85      1.47
 5 2000-04-01 CHE    CHF           Switzerla…        5.9       1.7 
 6 2000-04-01 CHL    CLP           Chile          1260       514   
 7 2000-04-01 CHN    CNY           China             9.9       8.28
 8 2000-04-01 CZE    CZK           Czech Rep…       54.4      39.1 
 9 2000-04-01 DNK    DKK           Denmark          24.8       8.04
10 2000-04-01 EUZ    EUR           Euro area         2.56      1.08
# … with 1,376 more rows, and 13 more variables:
#   dollar_price <dbl>, usd_raw <dbl>, eur_raw <dbl>,
#   gbp_raw <dbl>, jpy_raw <dbl>, cny_raw <dbl>, gdp_dollar <dbl>,
#   adj_price <dbl>, usd_adjusted <dbl>, eur_adjusted <dbl>,
#   gbp_adjusted <dbl>, jpy_adjusted <dbl>, cny_adjusted <dbl>

big_mac_gdp_data[, adj_price := lm(dollar_price ~ GDP_dollar) %>% predict, by = date]

big_mac |> 
  filter(date >= "2000-01-01",
         name %in% c("China", "Hong Kong", "Taiwan", "Japan", "Korean")) |> 
  ggplot(aes(x = date, y = local_price, color = name)) +
  geom_line(show.legend = FALSE) +
  facet_wrap(~name, scales = "free_y") +
  expand_limits(y = 0)

> capabilities()
       jpeg         png        tiff       tcltk         X11 
       TRUE        TRUE        TRUE        TRUE       FALSE 
       aqua    http/ftp     sockets      libxml        fifo 
       TRUE        TRUE        TRUE        TRUE        TRUE 
     cledit       iconv         NLS       Rprof     profmem 
       TRUE        TRUE        TRUE        TRUE        TRUE 
      cairo         ICU long.double     libcurl 
       TRUE        TRUE        TRUE        TRUE 

# library(tidyverse)
# library(janitor)
# library(zipcodeR)
# library(scales)

# library(usmap)

# library(sf)
# library(urbnmapr)


bu19 <- read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-05-11/broadband.csv', na = c("-", "", NA)) |>
  clean_names() |> 
  rename(state = st) |> 
  mutate(county_id = sprintf("%05d", county_id))

  state county_id county_name  broadband_availabil… broadband_usage
  <chr> <chr>     <chr>                       <dbl>           <dbl>
1 AL    01001     Autauga Cou…                 0.81            0.28
2 AL    01003     Baldwin Cou…                 0.88            0.3 
3 AL    01005     Barbour Cou…                 0.59            0.18
4 AL    01007     Bibb County                  0.29            0.07
5 AL    01009     Blount Coun…                 0.69            0.09
6 AL    01011     Bullock Cou…                 0.06            0.05

bu20 <- read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-05-11/broadband_zip.csv') |>
  clean_names() |>
  rename(state = st) |> 
  mutate(postal_code = sprintf("%05d", postal_code),
         county_id = sprintf("%05d", county_id)) |>

  # 分割
  left_join(zip_code_db |> select(zipcode, population),
            by = c(postal_code = "zipcode")) |> 
  mutate(q = population * broadband_usage) |> 
  group_by(county_id) |> 
  summarise(usage = sum(q) / sum(population))

# A tibble: 6 x 8
  state county_name   county_id postal_code broadband_usage
  <chr> <chr>         <chr>     <chr>                 <dbl>
1 SC    Abbeville     45001     29639                 0.948
2 SC    Abbeville     45001     29620                 0.398
3 SC    Abbeville     45001     29659                 0.206
4 SC    Abbeville     45001     29638                 0.369
5 SC    Abbeville     45001     29628                 0.221
6 LA    Acadia Parish 22001     70516                 0.032
# … with 3 more variables: error_range_mae <dbl>,
#   error_range_95_percent <dbl>, msd <dbl>

# A tibble: 6 x 2
  county_id usage
  <chr>     <dbl>
1 01001     0.377
2 01003     0.439
3 01005     0.292
4 01007     0.120
5 01009     0.197
6 01011     0.162

plot_usmap("counties", data = bu19 |> 
             rename(fips = county_id) |> 
             mutate(tier = cut(broadband_usage, 
                    breaks = c(0, 0.5, 1.0),
                    include.lowest = TRUE,
                    right = TRUE)),
           values = "tier",
           size = 0.1) +
  scale_fill_manual(values = c("#5ab4ac", "#f5f5f5"),
                    na.value = "#d8b365")

get_urbn_map("counties", sf = TRUE) |> 
  left_join(bu20, by = c("county_fips" = "county_id")) |> 
  mutate(tier = cut(usage, 
                    breaks = c(0, 0.5, 1.0),
                    include.lowest = TRUE,
                    right = TRUE)) |>
  ggplot() +
  geom_sf(aes(fill = tier), size = 0.1) +
  coord_sf(datum = NA) +
  scale_fill_manual(values = c("#5ab4ac", "#f5f5f5"),
                    na.value = "#d8b365") +
  theme_void() +
  theme(legend.position = "bottom")

gtag('event', 'view_item', {
  currency: 'USD',
  items: [
    {
      item_id: 'SKU_12345',
      item_name: 'jeggings',
      coupon: 'SUMMER_FUN',
      discount: 2.22,
      affiliation: 'Google Store',
      item_brand: 'Gucci',
      item_category: 'pants',
      item_variant: 'black',
      price: 9.99,
      currency: 'USD',
      quantity: 1,
    },
  ],
  value: 9.99,
});

v: 2
tid: G-ABCDEFG
gtm: 1234567
_p: 1039218207
sr: 2560x1440
ul: en-us
cid: 0123456789.9876543210
dl: https://placeless.net/blog/macos-dictionaries
dr: https://placeless.net/
dt: 《柯林斯双解》for macOS
sid: 1621444075
sct: 6
seg: 1
_s: 1

// Request Data

en=page_view
en=FCP&epn.value=1677&ep.metric_id=v1-1621444098204-8846853795459&epn.metric_value=1677&epn.metric_delta=1677&ep.metric_rating=good&up.color_mode=Dark&up.blocker=true

addEventListener('fetch', (event) => {
  // 可以追加一些请求过滤和回应的处理逻辑
  return event.respondWith(handleRequest(event.request));
});

async function readRequest(request) {
  const { url, headers } = request;
  const body = await request.text();
  const ga_url = url.replace(
    'https://你的域名/你的采集route',
    'https://www.google-analytics.com/g/collect'
  );
  const nq = {
    method: 'POST',
    headers: {
      Host: 'www.google-analytics.com',
      Origin: 'https://你的域名',
      'Cache-Control': 'max-age=0',
      'User-Agent': headers.get('user-agent'),
      Accept: headers.get('accept'),
      'Accept-Language': headers.get('accept-language'),
      'Content-Type': headers.get('content-type') || 'text/plain',
      Referer: headers.get('referer'),
    },
    body: body,
  };
  return new Request(nq);
}

async function handleRequest(event) {
  const newReq = await readRequest(event.request);
  event.waitUntil(fetch(newReq));
  return new Response(null, {
    status: 204,
    statusText: 'No Content',
  });
}

<script async src="/js/YOU-GA-SDK.js"></script>
<script>
  // function color_mode() {...}
  window.dataLayer = window.dataLayer || [];
  function gtag() { dataLayer.push(arguments); }
  gtag('js', new Date());

  gtag('config', 'GA4-Measurement-ID', { transport_url: 'https://YOUR-DOMAIN' });
  // gtag('set', 'user_properties', { color_mode: color_mode() });
</script>

library(tidyverse)
library(lubridate)

data %>%
  filter(created_at >= '2020-07-01',
         created_at < '2021-01-01',
         is.na(refund_state)) %>%
  select(member_no, union_no, created_at) %>% 
  mutate(created_at = date(created_at)) %>%
  group_by(member_no, union_no) %>%
  slice_max(created_at, n = 1) %>% 
  ungroup() %>%
  select(-union_no) %>%
  distinct() %>%
  group_by(member_no) %>%
  arrange(created_at) %>%
  mutate(
    n_days = n(),
    tier = case_when(
      n_days == 1 ~ "1d",
      n_days == 2 ~ "2d",
      n_days == 3 ~ "3d",
      n_days  > 3 ~ "3d+"),
    previous = lag(created_at),
    dd = interval(previous, created_at) %/% days(1))

data %>%
  filter(created_at >= '2020-07-01',
         created_at < '2021-01-01',
         is.na(refund_state)) %>%
  select(member_no, created_at) %>% 
  mutate(created_at = date(created_at)) %>%
  distinct()
  # ...

TZ="UTC"
RETICULATE_PYTHON="/usr/local/bin/python3"
R_MAX_VSIZE=16Gb

http_proxy=http://127.0.0.1:1091/
https_proxy=http://127.0.0.1:1091/

data %>%
  filter(created_at >= '2020-07-01',
         created_at < '2021-01-01',
         is.na(refund_state)) %>%
  select(member_no, created_at) %>% 
  mutate(created_at = date(created_at)) %>%
  distinct() %>% 
  group_by(member_no) %>%
  arrange(created_at) %>%
  mutate(
    n_days = n(),
    tier = case_when(
      n_days == 1 ~ "1d",
      n_days == 2 ~ "2d",
      n_days == 3 ~ "3d",
      n_days > 3 ~ "3d+"
    ),
    dd = c(0, diff(created_at)))

# data.table 原生 chain
dt[][
  ][
    ][
      ]

# 当然也可以使用 magrittr 的 pipe
# 但内容一多，括号里面的内容就不好分割了
dt %>%
  .[] %>%
  .[] %>%
  .[]

dt %>% # <----
  lazy_dt() %>% # <----
  filter(created_at >= '2020-07-01',
         created_at < '2021-01-01',
         is.na(refund_state)) %>%
  select(member_no, created_at) %>%
  mutate(created_at = date(created_at)) %>%
  distinct() %>%
  group_by(member_no) %>%
  arrange(created_at) %>%
  mutate(
    n_days = n(),
    tier = case_when(
      n_days == 1 ~ "1d",
      n_days == 2 ~ "2d",
      n_days == 3 ~ "3d",
      n_days > 3 ~ "3d+"
    ),
    dd = c(0, diff(created_at))
  ) %>%
  as.data.table() # <----

# system.time({...})

   user  system elapsed 
  2.915   0.037   2.098 

unique(`_DT2`[created_at >= "2020-07-01" & created_at < "2021-01-01" & 
    is.na(refund_state), .(member_no, created_at)][, `:=`(created_at = date(created_at))])[order(created_at)][, 
    `:=`(c("n_days", "tier", "dd"), {
        n_days <- .N
        tier <- fcase(n_days == 1, "1d", n_days == 2, "2d", n_days == 
            3, "3d", n_days > 3, "3d+")
        dd <- c(0, diff(created_at))
        .(n_days, tier, dd)
    }), by = .(member_no)]

Unit: seconds
 expr       min        lq      mean    median        uq      max neval
   dp 18.792854 19.282197 19.793766 19.602105 20.323187 20.96849     5
   dt  2.163451  2.220804  2.235009  2.234266  2.277006  2.27952     5

# 注释掉 case_when 和 diff 两处运算之后的对比
Unit: milliseconds
 expr       min        lq      mean    median       uq       max neval
   dp 1874.7782 1932.8641 2036.7339 2041.8302 2060.804 2273.3928     5
   dt  134.3752  138.8937  155.6884  141.1073  142.066  221.9998     5

async def g():
    # Pause here and come back to g() when f() is ready
    r = await f()
    return r

import json
from sys import argv
import pandas as pd
import asyncio
import aiohttp
import aiofiles
from tqdm.asyncio import tqdm

def read(file):
    pass

def getToken():
    pass
    
async def update(token, session, id, msg):
    async with session.put('url', json={"id": id, "msg": msg}) as response:
        text = await response.text()
        async with aiofiles.open(log_file, 'a') as f:
            await f.write(f'{text}\n')
        
async def main():
    data = read(argv[1])
    print(data.tail(5))
    print('Press Enter to continue...')
    token = getToken()
    
    async with aiohttp.ClientSession() as session:
        tasks = [update(token, session, row['id'], row['msg']) 
				 for _, row in data.iterrows()]
        results = [await t for t in tqdm.as_completed(tasks, leave=False)]
        return results
        
asyncio.run(main())

# Rime schema
# encoding: utf-8

schema:
  schema_id: placeless_flypy
  name: 小鹤双拼
  version: "0.1"
  author:
    - double pinyin layout by 鶴
    - Rime schema by 佛振 <chen.sst@gmail.com>
    - Mod by placeless
  description: |
    小鹤双拼自定义方案。

# 基本不用切换，因为不用扩展字符集，也不用 emoji
switches: 
  - name: full_shape
    reset: 0
    states: [ 半角, 全角 ]
  - name: ascii_punct
    states: [ 。，, ．， ]

# 只有简化字中文，所以
# 删掉了 ascii_composer
# 删掉了 ascii_segmentor
# 删掉了 filters
# selector 提前，方便;次选及'一选
engine:
  processors:
    - recognizer
    - key_binder
    - speller
    - selector
    - punctuator
    - navigator
    - express_editor
  segmentors:
    - matcher
    - abc_segmentor
    - punct_segmentor
    - fallback_segmentor
  translators:
    - punct_translator
    - script_translator

# 增加了 :./-=_+ 作为输入码
# 便于在打字中，不中断地输入 3.14、2:00-3:00 等
# 始码限制为：仅字母和数字
speller:
  alphabet: 'zyxwvutsrqponmlkjihgfedcba0987654321:,.!()/-=_+'
  initials: 'zyxwvutsrqponmlkjihgfedcba0987654321'
  algebra:
    - erase/^xx$/
    - derive/^([jqxy])u$/$1v/
    - derive/^([aoe])([ioun])$/$1$1$2/
    - xform/^([aoe])(ng)?$/$1$1$2/
    - xform/iu$/Q/
    - xform/(.)ei$/$1W/
    - xform/uan$/R/
    - xform/[uv]e$/T/
    - xform/un$/Y/
    - xform/^sh/U/
    - xform/^ch/I/
    - xform/^zh/V/
    - xform/uo$/O/
    - xform/ie$/P/
    - xform/i?ong$/S/
    - xform/ing$|uai$/K/
    - xform/(.)ai$/$1D/
    - xform/(.)en$/$1F/
    - xform/(.)eng$/$1G/
    - xform/[iu]ang$/L/
    - xform/(.)ang$/$1H/
    - xform/ian$/M/
    - xform/(.)an$/$1J/
    - xform/(.)ou$/$1Z/
    - xform/[iu]a$/X/
    - xform/iao$/N/
    - xform/(.)ao$/$1C/
    - xform/ui$/V/
    - xform/in$/B/
    - xlit/QWRTYUIOPSDFGHJKLZXCVBNM/qwrtyuiopsdfghjklzxcvbnm/

# 当前只有八股文简化字和袖珍简化字词库
# 其它词库，在 extended 中增加引用
# 繁体词库编译错误，需要修改前面的 switcher 和 filters
# 增加八股文简化字语言模型，优化长句输入
translator:
  dictionary: extended
  prism: placeless_flypy
  contextual_suggestions: true
  max_homophones: 7

grammar:
  language: zh-hans-t-essay-bgw

# 无前缀输入简单的预制表情和键盘符号
# 小鹤双拼没有 bq 和 kb 的编码
punctuator:
  import_preset: default
  symbols:
    "bq": [😂️, 😅️, 🎉, 🐂, 😱️, 👌, 😇️, 🙃️, 🤔️, 💊️, 💯️, 👍️, 🙈️, 💩️, 😈️ ]
    "kb": [⌘, ⌥, ⇧, ⌃, ⎋, ⇪, , ⌫, ⌦, ↩︎, ⏎, ↑, ↓, ←, →, ↖, ↘, ⇟, ⇞]
  half_shape:
    "\\" : "、"
    "#" : "#"
    "@" : "@"
    "~": "~"
    "/": "/"
    "'": {pair: ["「", "」"]}
    "[": "["
    "]": "]"
    "{": "{"
    "}": "}"
    "*": "*"
    "%": "%"
    "<" : ["<", "《"]
    ">" : [">", "》"]

recognizer:
  import_preset: default
  patterns:
    punct: "^(bq|kb)$"

# 💊️，只提供两个候选项
# 分号次选，空格或引号一选
# 首尾衔接，轮回翻页
menu:
  page_size: 2
  alternative_select_keys: "';"
  page_down_cycle: true

# Tab、Shift+Tab 上下翻页
# Control+p、Control+n 上下翻页
# Control+k = Esc，清除所有输入
key_binder:
  bindings:
    - { when: composing, accept: Tab,        send: Page_Down }
    - { when: composing, accept: Shift+Tab,  send: Page_Up }
    - { when: composing, accept: Control+k,  send: Escape }
    - { when: composing, accept: Control+p,  send: Page_Up }
    - { when: composing, accept: Control+n,  send: Page_Down }

# Rime dictionary
# encoding: utf-8

---
name: extended
version: "2020.12.30"
sort: by_weight
vocabulary: essay-zh-hans
use_preset_vocabulary: true

import_tables:
  - pinyin_simp
  - placeless_punct
...

# Rime dictionary
# encoding: utf-8
#

---
name: pinyin_simp
version: "0.2"
sort: by_weight
...

# zh_simp 精简 -> 8105

# Rime dictionary
# encoding: utf-8
#

---
name: placeless_punct
version: "0.2"
sort: by_weight
...

# 个人土方
.	.	1
,	,	1
:	:	1
/	/	1
-	-	1
_	_	1
=	=	1
+	+	1
0	0	1
1	1	1
2	2	1
3	3	1
4	4	1
5	5	1
6	6	1
7	7	1
8	8	1
9	9	1

patch:
  schema_list:
    - schema: placeless_flypy
  switcher/hotkeys:
    - "Control+grave"

patch:
  # 通知栏显示方式以及 ascii_mode 应用，与外观无关
  show_notifications_via_notification_center: true

  # 以下软件默认英文模式
  app_options: {}

  style:
    color_scheme: apathy
    horizontal: true # deprecated

  preset_color_schemes:
    apathy:
      author: "LIANG Hai | placeless"
      candidate_list_layout: linear # stacked | linear
      text_orientation: horizontal
      inline_preedit: true
      candidate_format: "%c'%@''"
      comment_text_color: 0x999999
      corner_radius: 5
      font_face: PingFangSC
      font_point: 17
      label_font_point: 17
      back_color: 0xFFFFFF
      text_color: 0x424242
      label_color: 0xFFFFFF
      hilited_candidate_back_color: 0xFFF0E4
      hilited_candidate_text_color: 0xEE6E00
      hilited_candidate_label_color: 0xFFF0E4
      name: "冷漠／Apathy"

- essay-zh-hans.txt
- pinyin_simp.dict.yaml
- zh-hans-t-essay-bgw.gram

bindkey -v
bindkey -M viins jj vi-cmd-mode
KEYTIMEOUT=25

inoremap jj <Esc>
:autocmd InsertEnter * set timeoutlen=200
:autocmd InsertLeave * set timeoutlen=1000

"vim.insertModeKeyBindings": [
     {
         "before": ["j", "j"],
         "after": ["<esc>"]
     }
]

from IPython import get_ipython
from prompt_toolkit.enums import DEFAULT_BUFFER
from prompt_toolkit.filters import HasFocus, ViInsertMode
from prompt_toolkit.key_binding.vi_state import InputMode

ip = get_ipython()

def switch_to_navigation_mode(event):
   vi_state = event.cli.vi_state
   vi_state.input_mode = InputMode.NAVIGATION

if getattr(ip, 'pt_app', None):
   registry = ip.pt_app.key_bindings
   registry.add_binding(u'j',u'j',
                        filter=(HasFocus(DEFAULT_BUFFER)
                                 & ViInsertMode()))(switch_to_navigation_mode)

"imap": 
  [
      ["jj", "<Esc>"],
  ]

# for f in *.csv; do 7z a -tgzip "$f.zip" "$f"; done

for f in *.csv; do gzip $i; done

$ file *.zip
a.zip: gzip compressed data, was "affin...
b.zip: Zip archive data, at least v?[0x314] to extract
c.zip: 7-zip archive data, version 0.4

$ zipinfo b.zip
Archive:  b.zip
Zip file size: 1659 bytes, number of entries: 1
-rw-r--r--  6.3 unx     9898 bx defN 20-Apr-23 00:09 affinity_categories.csv
1 file, 9898 bytes uncompressed, 1479 bytes compressed:  85.1%

gzcat a.zip | head -5
# or 
gunzip -c a.zip | less

unzip -c b.zip | less

7zcat c.zip | less

zgrep Love a.gz | head -3
zipgrep Love b.zip | less