add CSV support, adhoc scripts and adhoc analysis

.gitignore

@@ -1,4 +1,9 @@
 .idea
 data
 dist
-**/__pycache__
+**/__pycache__
+.Rproj.user
+.RData
+*.html
+*.log
+*.csv

README.md

@@ -2,27 +2,40 @@
 
 ## Installation
 
-```
-pip install logtools
+```sh
+pip install pip install git+https://github.com/gmega/logtools.git
 ```
 
 ## Usage
 
 ### Merge by Timestamp
 
-```
+```sh
 log-merge log1.log log2.log
 ```
 
 ### Merge by Timestamp Showing Aliases Instead of File Name
 
-```
+```sh
 log-merge log1.log log2.log --aliases bootstrap codex21
 ```
 
 ### Merge and Filter by Timestamp
 
-```
+```sh
 # If no timezone is provided, assumes UTC
 log-merge log1.log log2.log --from 2021-01-01T00:00:00 --to 2021-01-02T00:00:00
 ```
+
+### Transform Raw Logs into CSV
+
+```sh
+cat ./log1.log | log-to-csv
+```
+
+### Transform Raw Logs into CSV, Extracting Topics Into Column
+
+```sh
+cat ./log1.log | log-to-csv --extract-fields topics
+```
+

adhoc/identify_uploads.py

@@ -0,0 +1,18 @@
+"""Ad-hoc script which tags uploads with a sequential number."""
+import sys
+
+uploading = False
+upload_no = 0
+for line in sys.stdin:
+    if 'Handling file upload' in line:
+        upload_no += 1
+        uploading = True
+
+    if uploading:
+        line = line.strip()
+        line = line.rsplit(' ', maxsplit=1)
+        line = ' '.join([line[0], f'upload={upload_no}', line[1]])
+        print(line)
+
+    if 'Uploaded file' in line:
+        uploading = False

analysis/.Rhistory

@@ -0,0 +1,512 @@
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1:2], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 13) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[c(-1,-2)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 13) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[c(-1,-2,-3)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 13) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[c(-1,-2,-4)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 13) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[c(-1,-2,-3)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 13) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-(1:3)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-(1:4)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-(1:5)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-(1:10)], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+c(1,2,3,4)
+c(1,2,3,4)[-1]
+c(1,2,3,4)[-2]
+c(1,2,3,4)[-(1:2)]
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval, probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds'))
+ggplot(
+interlog_intervals |>
+filter(upload == 18) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 1) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 2) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 3) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 4) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 5) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 6) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 7) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 8) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 7) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 8) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 9) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 10) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 11) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 12) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 13) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 14) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 15) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 16) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 17) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()
+ggplot(
+interlog_intervals |>
+filter(upload == 18
+) |>
+mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+group_by(bucket) |>
+mutate(
+mean_interval = mean(log_interval),
+p_70 = quantile(log_interval[-1], probs = c(0.95))
+) |>
+ungroup()
+) +
+geom_point(aes(x = timestamp, y = log_interval)) +
+geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) +
+theme_minimal()

analysis/analysis.Rmd

@@ -0,0 +1,152 @@
+---
+title: "R Notebook"
+output: html_notebook
+---
+
+```{r}
+library(tidyverse)
+library(lubridate)
+```
+
+```{r}
+uploads <- read_csv('./codex-continuous-tests-0codex3-5-77bdb95dc7-j7f46_codex3-5-uploads.csv')
+```
+
+
+```{r}
+durations <- uploads |> 
+  arrange(count) |> 
+  group_by(upload) |> 
+  summarise(
+    start = timestamp[1],
+    end = timestamp[n()],
+  ) |> 
+  mutate(duration = end - start)
+```
+
+How long are uploads taking?
+
+```{r}
+ggplot(durations, aes(x = upload, y = duration)) + 
+  geom_point() + 
+  geom_line() +
+  ylab('upload duration') + 
+  xlab('upload number') + 
+  theme_minimal()
+```
+Are all uploads completing?
+
+```{r}
+uploads |> 
+  filter(message == 'Got data from stream') |> 
+  group_by(upload) |> 
+  count(name = 'blocks')
+```
+
+Does the end of the upload coincide with the last chunk that gets stored?
+
+```{r}
+uploads |> 
+  filter(grepl('Got data from stream', message)) |> 
+  group_by(upload) |>
+  summarise(
+    last_store = max(timestamp)
+  ) |> 
+  inner_join(durations, by='upload')
+```
+
+```{r}
+durations
+```
+
+```{r}
+uploads |> filter(grepl('Exception', message)) |> group_by(message) |> count() |> arrange(n)
+```
+
+```{r}
+uploads |> filter(upload == 18) |> group_by(message) |> count() |> arrange(n)
+```
+
+```{r}
+uploads |> filter(upload == 17) |> group_by(message)  |> count() |> arrange(n)
+```
+
+```{r}
+messages <- uploads |> group_by(message) |> count() |> filter(n > 100) |> pull(message)
+```
+
+
+```{r fig.height = 10}
+uploads |> filter(message %in% messages) |> group_by(upload, message) |> count() %>% {
+  ggplot(.) + 
+    geom_point(aes(x = message, y = n, color = as.factor(upload))) + theme_minimal() + theme(axis.text.x = element_text(angle = 45, hjust=1)) +
+    ylab('count') + 
+    scale_color_manual(values=c('18'='red'))
+}
+```
+
+
+```{r}
+interlog_intervals <- uploads |> 
+  group_by(upload) |> 
+  arrange(timestamp) |> 
+  mutate(log_interval = as.numeric(timestamp - lag(timestamp))) |>
+  ungroup()
+```
+
+
+```{r}
+interlog_intervals |> 
+  group_by(upload) |>
+  summarise(
+    mean_li = mean(log_interval, na.rm=TRUE),
+    median_li = median(log_interval, na.rm=TRUE),
+    max_li = max(log_interval, na.rm=TRUE),
+  ) |>
+  pivot_longer(-upload) %>% {
+    ggplot(.) + 
+      geom_line(aes(x = upload, y = value, col = name)) +
+      scale_y_log10() + 
+      theme_minimal() +
+      ylab('duration (logscale, seconds)')
+  }
+```
+
+```{r}
+interlog_intervals |> group_by(upload) |> count() |> arrange(desc(n))
+```
+
+
+```{r fig.height=5}
+interlog_intervals |> 
+  group_by(upload) |>
+  arrange(log_interval) |>
+  mutate(rank = seq_along(log_interval)) |> ungroup() %>% {
+    ggplot(.) + 
+      geom_point(aes(x = rank, y = log_interval, col = as.factor(upload))) + 
+      theme_minimal() +
+      xlab('rank') + 
+      ylab('time between two consecutive log messages') +
+      guides(col = guide_legend(title = 'upload #'))
+  }
+```
+
+```{r}
+ggplot(
+  interlog_intervals |> 
+    filter(upload == 18
+           ) |>
+    mutate(bucket = floor_date(timestamp, unit = '5 seconds')) |>
+    group_by(bucket) |>
+    mutate(
+      mean_interval = mean(log_interval),
+      p_70 = quantile(log_interval[-1], probs = c(0.95))
+    ) |> 
+    ungroup()
+  ) + 
+  geom_point(aes(x = timestamp, y = log_interval)) +
+  geom_line(aes(x = bucket, y = mean_interval), col = 'red', lwd = 2) +
+  geom_line(aes(x = bucket, y = p_70), col = 'orange', lwd = 2) + 
+  theme_minimal()
+```
+

analysis/analysis.Rproj

@@ -0,0 +1,13 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX

logtools/cli/merge.py

@@ -57,8 +57,9 @@ def _assign_colors(names: Dict[str, str]) -> Dict[str, str]:
 def _filtering_predicate(args):
     if args.from_ or args.to:
         return timestamp_range(
-            _ensure_utc(tsparser.parse(args.from_)),
-            _ensure_utc(tsparser.parse(args.to))
+            _ensure_utc(args.from_) if args.from_ is not None else datetime(
+                year=1980, month=1, day=1, hour=0, minute=0, second=0, tzinfo=pytz.UTC),
+            _ensure_utc(args.to) if args.to is not None else datetime.utcnow().replace(tzinfo=pytz.UTC)
         )
 
     return lambda x: True

logtools/cli/to_csv.py

@@ -0,0 +1,38 @@
+"""Parses a log stream, possibly pre-filtered and/or merged, into a CSV file. Allows chronicles topics to be
+extracted into their own columns."""
+import sys
+from csv import DictWriter
+
+from traitlets.config.loader import ArgumentParser
+
+from logtools.log.sources.stream_log_source import StreamLogSource
+
+
+def to_csv(args):
+    fields = args.extract_fields
+    writer = DictWriter(sys.stdout,
+                        fieldnames=['timestamp', 'line_number', 'level', 'fields', 'count', 'message'] + fields)
+    writer.writeheader()
+    for line in StreamLogSource(sys.stdin):
+        line_fields = {field: line.fields.get(field, 'NA') for field in fields}
+        writer.writerow({
+            'timestamp': line.timestamp.isoformat(),
+            'line_number': line.location.line_number,
+            'level': line.level.value,
+            'fields': line.topics,
+            'count': line.count,
+            'message': line.message,
+            **line_fields,
+        })
+
+
+def main():
+    argparse = ArgumentParser()
+    argparse.add_argument('--extract-fields', nargs='+', default=[],
+                          help='Extract chronicles topics into CSV columns')
+
+    to_csv(argparse.parse_args())
+
+
+if __name__ == '__main__':
+    main()

logtools/log/composers.py

@@ -1,165 +0,0 @@
-# import abc
-# import re
-# from abc import abstractmethod
-# from datetime import datetime
-# from heapq import heapify, heappop, heappush
-# from pathlib import Path
-# from typing import TypedDict, Iterable, Union, Generator, Optional, Iterator, List
-#
-# from parse.utils import group_match
-#
-# class LogIterator(abc.ABC, Iterator[LogLine]):
-#     @abstractmethod
-#     def peek(self) -> Optional[LogLine]:
-#         ...
-#
-#     def context(self) -> str:
-#         ...
-#
-#
-# class SingleLogIterator(LogIterator):
-#
-#     def __init__(
-#             self,
-#             path: Path,
-#             alias: str,
-#             from_ts: Optional[datetime] = None,
-#             to_ts: Optional[datetime] = None,
-#             parse_datetime=False
-#     ):
-#         self.path = path
-#         self.line_number = 0
-#         self.parse_datetime = parse_datetime
-#         self.alias = alias
-#
-#         # If from_ts or to_ts is specified, then timestamp parsing is mandatory.
-#         self.parse_datetime = self.parse_datetime or (from_ts is not None or to_ts is not None)
-#         self.from_ts = from_ts
-#         self.to_ts = to_ts
-#
-#         self.inner_iterator = self._iterator()
-#         self.look_ahead = next(self.inner_iterator, None)
-#
-#     def __next__(self) -> LogLine:
-#         next_element = self.look_ahead if self.look_ahead is not None else next(self.inner_iterator)
-#         self.look_ahead = next(self.inner_iterator, None)
-#         return next_element
-#
-#     def __iter__(self):
-#         return self
-#
-#     def __lt__(self, other):
-#         return self.latest_timestamp() < other.latest_timestamp()
-#
-#     def __le__(self, other):
-#         return self.latest_timestamp() <= other.latest_timestamp()
-#
-#     def _iterator(self) -> Generator[LogLine, None, None]:
-#         with self.path.open() as f:
-#             for line in f:
-#                 self.line_number += 1
-#                 contents = group_match(line, LOG_LINE)
-#                 if not contents:
-#                     continue
-#
-#                 line = LogLine(
-#                     parent=self,
-#                     log=self.alias,
-#                     raw=line,
-#                     line_number=self.line_number,
-#                     timestamp=(datetime.fromisoformat(contents['timestamp']) if self.parse_datetime
-#                                else contents['timestamp']),
-#                     message=contents['message'],
-#                 )
-#
-#                 if self.should_accept(line):
-#                     yield line
-#
-#     def should_accept(self, line: LogLine) -> bool:
-#         timestamp = line['timestamp']
-#         if self.from_ts is not None and timestamp <= self.from_ts:
-#             return False
-#
-#         if self.to_ts is not None and timestamp >= self.to_ts:
-#             return False
-#
-#         return True
-#
-#     def peek(self) -> Optional[LogLine]:
-#         return self.look_ahead
-#
-#     def latest_timestamp(self) -> Optional[datetime]:
-#         return self.peek()['timestamp'] if self.peek() is not None else None
-#
-#     def context(self) -> str:
-#         return f'{self.path}:{self.line_number}'
-#
-#
-# def _exclude_empty(logs: Iterable[LogIterator]):
-#     return [log for log in logs if log.peek() is not None]
-#
-#
-# class CollatingLogIterator(LogIterator):
-#
-#     def __init__(self, logs: List[SingleLogIterator]):
-#         self.logs = _exclude_empty(logs)
-#
-#     def __iter__(self):
-#         return self
-#
-#     def __next__(self):
-#         if not self.logs:
-#             raise StopIteration()
-#
-#         log = self.logs[0]
-#         value = next(log)
-#         if log.peek() is None:
-#             self.logs.pop(0)
-#         return value
-#
-#     def peek(self) -> Optional[LogLine]:
-#         if not self.logs:
-#             return None
-#
-#         return self.logs[0].peek()
-#
-#     def context(self) -> str:
-#         if not self.logs:
-#             raise Exception('Undefined context.')
-#
-#         return self.logs[0].context()
-#
-#
-# class MergingLogIterator(LogIterator):
-#     def __init__(self, logs: List[SingleLogIterator]):
-#         self.logs = _exclude_empty(logs)
-#         heapify(self.logs)
-#
-#     def __iter__(self):
-#         return self
-#
-#     def __next__(self) -> LogLine:
-#         if not self.logs:
-#             raise StopIteration()
-#
-#         # by construction, we can't have any empty iterators at this point, so the call to next always succeeds.
-#         log = heappop(self.logs)
-#         value = next(log)
-#
-#         # if the iterator still has stuff in it...
-#         if log.peek() is not None:
-#             heappush(self.logs, log)
-#
-#         return value
-#
-#     def peek(self) -> Optional[LogLine]:
-#         if not self.logs:
-#             return None
-#
-#         return self.logs[0].peek()
-#
-#     def context(self) -> str:
-#         if not self.logs:
-#             raise Exception('Undefined context.')
-#
-#         return self.logs[0].context()

logtools/log/log_line.py

@@ -2,7 +2,9 @@ import re
 from dataclasses import dataclass
 from datetime import datetime
 from enum import Enum
-from typing import Union, Self, Optional
+from typing import Union, Optional
+
+TOPICS = re.compile(r'(?P<key>\w+)=(?P<value>"[\w\s]+"|\S+)')
 
 
 class LogLevel(Enum):
@@ -10,35 +12,20 @@ class LogLevel(Enum):
     debug = 'DBG'
     info = 'INF'
     error = 'ERR'
-
-
-LOG_LINE = re.compile(
-    r'(?P<line_type>\w{3}) (?P<timestamp>\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}.\d{3}\+\d{2}:\d{2}) (?P<message>.*) '
-    r'count=(?P<count>\d+)$'
-)
+    warning = 'WRN'
+    note = 'NOT'
 
 
 @dataclass
 class LogLine:
     raw: str
     level: LogLevel
-    line_number: int
     timestamp: Union[str, datetime]
     message: str
+    topics: str
     count: Optional[int]
 
-    @classmethod
-    def from_str(cls, source: str, parse_datetime: bool = False) -> Self:
-        parsed = LOG_LINE.search(source)
-        if not parsed:
-            raise ValueError(f'Could not parse log line: {source}')
-
-        return cls(
-            raw=source,
-            level=LogLevel(parsed['line_type'].upper()),
-            line_number=0,
-            timestamp=(datetime.fromisoformat(parsed['timestamp']) if parse_datetime
-                       else parsed['timestamp']),
-            message=parsed['message'],
-            count=int(parsed['count']) if parsed['count'] else None,
-        )
+    @property
+    def fields(self):
+        fields = TOPICS.findall(self.topics)
+        return {key: value for key, value in fields} if fields else {}

logtools/log/sources/file_log_source.py

@@ -1,30 +1,25 @@
-import sys
 from dataclasses import dataclass
 from pathlib import Path
 
-from logtools.log.sources.log_source import TrackedLogLine, LogSource
+from logtools.log.sources.log_parsers import LineNumberLocation
+from logtools.log.sources.stream_log_source import StreamLogSource
 
 
 @dataclass
-class FileLineLocation:
+class FileLineLocation(LineNumberLocation):
     path: Path
-    line_number: int
 
 
-class FileLogSource(LogSource[TrackedLogLine[FileLineLocation]]):
+class FileLogSource(StreamLogSource):
     def __init__(self, path: Path, parse_datetime=True):
         self.path = path
-        self.parse_datetime = parse_datetime
+        super().__init__(self.path.open(encoding='utf-8'), parse_datetime=parse_datetime)
 
     def __iter__(self):
-        with self.path.open(encoding='utf-8') as f:
-            for line_number, line in enumerate(f, start=1):
-                try:
-                    parsed = TrackedLogLine.from_str(line, parse_datetime=True)
-                    parsed.location = FileLineLocation(self.path, line_number)
+        try:
+            yield from super().__iter__()
+        finally:
+            self.stream.close()
 
-                    yield parsed
-                except ValueError:
-                    # FIXME we should probably relax parsing restrictions and output
-                    #   these too but for now just skip it.
-                    print(f'Skip unparseable line: {line}', file=sys.stderr)
+    def _location(self, line_number: int) -> LineNumberLocation:
+        return FileLineLocation(path=self.path, line_number=line_number)

logtools/log/sources/filtered_source.py

@@ -17,6 +17,6 @@ class FilteredSource(LogSource[TrackedLogLine[TLocation]]):
 
 def timestamp_range(start: datetime, end: datetime):
     def predicate(line: TrackedLogLine[TLocation]):
-        return start <= line.timestamp <= end
+        return start <= line.timestamp <= end  # type: ignore
 
     return predicate

logtools/log/sources/log_parsers.py

@@ -0,0 +1,76 @@
+import re
+import sys
+from csv import DictReader
+from dataclasses import dataclass
+
+from typing import Callable, TextIO, Optional, cast
+
+from dateutil import parser as tsparser
+
+from logtools.log.log_line import LogLevel
+from logtools.log.sources.log_source import TrackedLogLine, LogSource
+
+
+@dataclass
+class LineNumberLocation:
+    line_number: int
+
+
+"""A :class:`LogParser` is a function that takes a raw text stream and returns a :class:`LogSource`, which in turn
+is an iterable of parsed lines."""
+LogParser = Callable[[TextIO], LogSource[LineNumberLocation]]
+
+LOG_LINE = re.compile(
+    r'(?P<line_type>\w{3}) (?P<timestamp>\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}.\d{3}\+\d{2}:\d{2}) (?P<message>.*) '
+    r'count=(?P<count>\d+)$'
+)
+
+TOPICS = re.compile(r'((\w+=("[\w\s]+"|\S+) )+)?\w+=("[\w\s]+"|\S+)$')
+
+
+def parse_raw(line: str, parse_datetime: bool = True) -> Optional[TrackedLogLine[LineNumberLocation]]:
+    parsed = LOG_LINE.search(line)
+    topics = TOPICS.search(parsed['message'])
+    if not parsed or not topics:
+        return None
+
+    return TrackedLogLine(
+        raw=line,
+        level=LogLevel(parsed['line_type'].upper()),
+        timestamp=(tsparser.parse(parsed['timestamp']) if parse_datetime
+                   else parsed['timestamp']),
+        message=parsed['message'][:topics.start() - 1].strip(),
+        count=int(parsed['count']) if parsed['count'] else None,
+        topics=topics.group()
+    )
+
+
+def raw_parser(stream: TextIO, parse_datetime=True) -> LogSource:
+    for line_number, line in enumerate(stream, start=1):
+        parsed = parse_raw(line, parse_datetime=parse_datetime)
+        if not parsed:
+            # FIXME we should probably relax parsing restrictions and output
+            #   these too but for now just skip it.
+            print(f'Skip unparseable line: {line}', file=sys.stderr)
+            continue
+
+        yield parsed
+
+
+def csv_parser(stream: TextIO, parse_datetime=True) -> LogSource:
+    for line_number, line in enumerate(DictReader(stream), start=1):
+        try:
+            line = TrackedLogLine(
+                raw=line['message'],  # FIXME this is NOT the raw line...
+                timestamp=line['timestamp'],
+                message=line['message'],
+                count=int(line['count']) if line['count'] else None,
+                topics=line['topics'],
+                level=LogLevel[line['level']],
+            )
+
+            if parse_datetime:
+                line.timestamp = tsparser.parse(cast(str, line.timestamp))
+            yield line
+        except ValueError:
+            print(f'Skip unparseable line: {line}', file=sys.stderr)

logtools/log/sources/stream_log_source.py

@@ -0,0 +1,19 @@
+from typing import TextIO
+
+from logtools.log.sources.log_parsers import raw_parser, LineNumberLocation, LogParser
+from logtools.log.sources.log_source import LogSource, TrackedLogLine
+
+
+class StreamLogSource(LogSource[TrackedLogLine[LineNumberLocation]]):
+    def __init__(self, stream: TextIO, parse_datetime=True, log_format: LogParser = raw_parser):
+        self.stream = stream
+        self.format = log_format
+        self.parse_datetime = parse_datetime
+
+    def __iter__(self):
+        for line_number, line in enumerate(self.format(self.stream), start=1):
+            line.location = self._location(line_number)
+            yield line
+
+    def _location(self, line_number: int) -> LineNumberLocation:
+        return LineNumberLocation(line_number)

logtools/log/sources/tests/string_log_source.py

@@ -1,22 +1,19 @@
 from dataclasses import dataclass
+from io import StringIO
 
-from logtools.log.sources.log_source import LogSource, TrackedLogLine
+from logtools.log.sources.log_parsers import LogParser
+from logtools.log.sources.stream_log_source import StreamLogSource, LineNumberLocation, raw_parser
 
 
 @dataclass
-class ParseLocation:
+class ParseLocation(LineNumberLocation):
     name: str
-    number: int
 
 
-class StringLogSource(LogSource[TrackedLogLine[ParseLocation]]):
-    def __init__(self, name: str, lines: str):
+class StringLogSource(StreamLogSource):
+    def __init__(self, name: str, lines: str, log_format: LogParser = raw_parser):
         self.name = name
-        self.lines = lines
+        super().__init__(stream=StringIO(lines), log_format=log_format)
 
-    def __iter__(self):
-        for line_number, line in enumerate(self.lines.splitlines(), start=1):
-            parsed = TrackedLogLine.from_str(line, parse_datetime=True)
-            parsed.location = ParseLocation(self.name, line_number)
-
-            yield parsed
+    def _location(self, line_number: int) -> LineNumberLocation:
+        return ParseLocation(name=self.name, line_number=line_number)

logtools/log/sources/tests/test_collating_source.py

@@ -22,7 +22,7 @@ def test_should_collate_lines_from_log_sources():
     )
 
     collated = CollatingSource(log1, log2)
-    entries = [(line.location.name, line.location.number, line.count) for line in collated]
+    entries = [(line.location.name, line.location.line_number, line.count) for line in collated]
     assert entries == [
         ('log1', 1, 1),
         ('log1', 2, 2),

logtools/log/sources/tests/test_filtered_source.py

@@ -2,6 +2,7 @@ from dateutil import parser
 
 from logtools.log.log_line import LogLine
 from logtools.log.sources.filtered_source import FilteredSource, timestamp_range
+from logtools.log.sources.log_parsers import parse_raw
 from logtools.log.sources.tests.string_log_source import StringLogSource
 
 
@@ -32,7 +33,7 @@ def test_should_generate_correct_datetime_range_predicate():
     matches = timestamp_range(start=parser.parse('2023-10-16 22:29:24.597+00:00'),
                               end=parser.parse('2023-10-18 20:29:25.597+00:00'))
 
-    lines = [LogLine.from_str(line, parse_datetime=True) for line in raw_lines]
+    lines = [parse_raw(line, parse_datetime=True) for line in raw_lines]
     filtered = [line.count for line in lines if matches(line)]
 
     assert filtered == [2, 3]

logtools/log/sources/tests/test_log_parsers.py

@@ -0,0 +1,19 @@
+from datetime import datetime
+
+import pytz
+
+from logtools.log.log_line import LogLevel
+from logtools.log.sources.log_parsers import parse_raw
+
+
+def test_raw_parser_should_parse_logline_from_string():
+    line = parse_raw('TRC 2023-10-16 17:28:46.579+00:00 Sending want list to peer                  '
+                     'topics="codex blockexcnetwork" tid=1 peer=16U*7mogoM '
+                     'type=WantBlock items=1 count=870781', parse_datetime=True)
+
+    assert line.level == LogLevel.trace
+    assert line.timestamp == datetime(2023, 10, 16, 17, 28, 46,
+                                      579000, tzinfo=pytz.utc)
+    assert line.message == 'Sending want list to peer'
+    assert line.topics == 'topics="codex blockexcnetwork" tid=1 peer=16U*7mogoM type=WantBlock items=1'
+    assert line.count == 870781

logtools/log/sources/tests/test_ordered_source.py

@@ -5,15 +5,12 @@ from logtools.log.sources.tests.string_log_source import StringLogSource
 
 
 def test_should_order_sources_by_lookahead_timestamp():
-    contents = StringLogSource(
-        name='log1',
-        lines="""TRC 2023-10-16 20:29:24.595+00:00 Advertising block        topics="codex discoveryengine" count=1
+    lines = """TRC 2023-10-16 20:29:24.595+00:00 Advertising block        topics="codex discoveryengine" count=1
               TRC 2023-10-16 20:29:24.597+00:00 Provided to nodes           topics="codex discovery" tid=1 count=2
               TRC 2023-10-16 20:29:24.646+00:00 Retrieved record from repo  topics="codex repostore" count=3"""
-    )
 
-    log1 = OrderedSource(contents)
-    log2 = OrderedSource(contents)
+    log1 = OrderedSource(StringLogSource(name='log1', lines=lines))
+    log2 = OrderedSource(StringLogSource(name='log2', lines=lines))
 
     next(log1)
     assert log2 < log1
@@ -24,19 +21,15 @@ def test_should_order_sources_by_lookahead_timestamp():
 
 
 def test_should_raise_error_if_comparing_empty_sources():
-    contents = StringLogSource(
-        name='log1',
-        lines="""TRC 2023-10-16 20:29:24.595+00:00 Advertising block        topics="codex discoveryengine" count=1
+    lines = """TRC 2023-10-16 20:29:24.595+00:00 Advertising block        topics="codex discoveryengine" count=1
               TRC 2023-10-16 20:29:24.597+00:00 Provided to nodes           topics="codex discovery" tid=1 count=2
               TRC 2023-10-16 20:29:24.646+00:00 Retrieved record from repo  topics="codex repostore" count=3"""
-    )
 
-    log1 = OrderedSource(contents)
-    log2 = OrderedSource(contents)
+    log1 = OrderedSource(StringLogSource(name='log1', lines=lines))
+    log2 = OrderedSource(StringLogSource(name='log2', lines=lines))
 
     for _ in log1:
         ...
 
     with pytest.raises(ValueError):
         _ = log1 < log2
-

logtools/log/tests/test_logline.py

@@ -1,16 +1,19 @@
-from datetime import datetime
-
-import pytz
-
 from logtools.log.log_line import LogLine, LogLevel
 
 
-def test_should_parse_logline_from_string():
-    line = LogLine.from_str('TRC 2023-10-16 17:28:46.579+00:00 Sending want list to peer                  '
-                            'topics="codex blockexcnetwork" tid=1 peer=16U*7mogoM '
-                            'type=WantBlock items=1 count=870781', parse_datetime=True)
+def test_should_parse_chronicles_fields():
+    line = LogLine(message='Sending want list to peer',
+                   topics='topics="codex blockexcnetwork" tid=1 peer=16U*7mogoM '
+                          'type=WantBlock items=1',
+                   timestamp='',
+                   count=0,
+                   raw='',
+                   level=LogLevel.trace)
 
-    assert line.level == LogLevel.trace
-    assert line.timestamp == datetime(2023, 10, 16, 17, 28, 46,
-                                      579000, tzinfo=pytz.utc)
-    assert line.count == 870781
+    assert line.fields == {
+        'topics': '"codex blockexcnetwork"',
+        'tid': '1',
+        'peer': '16U*7mogoM',
+        'type': 'WantBlock',
+        'items': '1',
+    }

pyproject.toml

@@ -24,4 +24,5 @@ requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
 
 [tool.poetry.scripts]
-log-merge = 'logtools.cli.merge:main'
+log-merge = 'logtools.cli.merge:main'
+log-to-csv = 'logtools.cli.to_csv:main'