You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Respond: user:test bot:rose ip: (~introductions) 1 where do you live ==> I'm not from around here. When:Jan15'16-17:03:46
17
15
Why:~no_task.0.0.~control.21.0=MAIN
18
16
```
19
-
A "start" line shows conversation initiation. After Start, the line contains the user name, the bot name,
20
-
the IP address (using a default configuration). Then the current value of the random seed and what topic
21
-
it ended in in parens and the volley count. That completes the input side. The ==> indicates the
22
-
upcoming outputs side. You see the startup message it issued (eg _Hello, I'm Rose._). Then a
23
-
datestamp of when this was issued, followed by the CS engine version, the date stamps when build0,
24
-
build1, and build2 were created. `F:` and `P:` are specially volley markers for reconstructing
25
-
conversations. And Why tells you the rule tags of the rule and possibly the reuse prior rule that
26
-
immediately generated the output.
17
+
A `start` line shows conversation initiation. After Start, the line contains the user name, the bot name,
18
+
the IP address (using a default configuration).
19
+
Then the current value of the random seed and what topic it ended in in parens and the volley count. That completes the input side.
20
+
21
+
The `==>` indicates the upcoming outputs side. You see the startup message it issued (eg _Hello, I'm Rose._).
22
+
Then a datestamp of when this was issued, followed by the CS engine version, the date stamps when build0,
23
+
build1, and build2 were created.
24
+
25
+
`F:` and `P:` are specially volley markers for reconstructing conversations.
26
+
And Why tells you the rule tags of the rule and possibly the reuse prior rule that immediately generated the output.
27
27
The respond log entry shows you similar information, the only difference being that immediately
28
28
before the `==>` is the actual input from the user.
29
29
@@ -99,23 +99,23 @@ You can do all topics in a file by naming the file name instead of the topicname
99
99
path, just the actual name of the file.
100
100
101
101
102
-
# Views over User Logs - :trim
102
+
# Views over User Logs - `:trim`
103
103
104
104
If you get a lot of user logs (say thousands), reading through them becomes a chore. The logs have a
105
105
bunch of excess information and are in a bunch of different files. This is where `:trim` comes in, making
106
106
it easier to see things. `:trim` assumes all files in the LOGS directory are user logs and will process them
107
-
in some manner. It will normally put its output in `TMP/tmp.txt`. Your first argument to trim can also be
108
-
just the name of a user (whose log file will be in `USERS/log-xxx.txt`) or if that doesn't exist then it is
109
-
the directory to use, or you can use the name of a log file within the USERs directory (the name should
110
-
begin "log-" and not include the directory and need not include the .txt suffix). E.g.,
107
+
in some manner. It will normally put its output in `TMP/tmp.txt`.
108
+
109
+
Your first argument to trim can also be just the name of a user (whose log file will be in `USERS/log-xxx.txt`)
110
+
or if that doesn't exist then it is the directory to use, or you can use the name of a log file within the USERs directory (the name should begin `log-` and not include the directory and need not include the `.txt` suffix). E.g.,
111
111
```
112
112
:trim c:\FULLLOGS 6
113
113
```
114
114
a directory to read all files within
115
115
```
116
116
:trim log-bob 6
117
117
```
118
-
named log file with .txt defaulted
118
+
named log file with `.txt` defaulted
119
119
```
120
120
:trim bob 6
121
121
```
@@ -125,26 +125,24 @@ user with logfile log-bob.txt in USERS
125
125
```
126
126
Trim will read every file and generate output depending on the integer code given it. The codes are:
127
127
128
-
| n | description |
129
-
| ---- | ----------- |
130
-
| `0` | puts the what the user said, followed by what the chatbot said, on single lines, removing all the excess junk.
131
-
| `1` | similar to 0, but puts what the chatbot said first, and what the user said after. This is useful for seeing all the responses users have made and can be aggregated to figure out what clever rejoinders you might want.
132
-
| `2` | similar to 0 (user first), but puts the name of the topic the chatbot ended in before either. You can see the flow of topics better with this view.
133
-
| `3` | similar to 2 (topic shown), but puts what the chatbot said before the user.
134
-
| `4` | puts the user and chatbot on separate lines, indenting the chatbots line. Easier to read.
135
-
| `5` | similar to 3, but indents the user instead of the chatbot.
136
-
| `6` | only lists the users inputs. This is good for creating a file that can recreate a user's experience, if you want to recreate it for debugging or regression.
137
-
| `7` | display rule responsible for output. Analogous to :why, it shows the rule tag, the sample input comment if there is one, the rule type and pattern, the input from the user and the output from the chatbot. If the rule doing the output was the target of a local ^reuse (same topic), then the data about the
138
-
rule comes from the calling rule, not the output rule.
139
-
| `8` | puts the user and chatbot on separate lines, indenting the chatbots line and prefixes it with the topic generating the response. Easier to read and debug.
140
-
| `11` | puts the timestamp and user on first line and and chatbot on second line,indenting the chatbots line.
128
+
| n | description
129
+
| --- | -----------
130
+
| `0` | puts the what the user said, followed by what the chatbot said, on single lines, removing all the excess junk.
131
+
| `1` | similar to 0, but puts what the chatbot said first, and what the user said after. This is useful for seeing all the responses users have made and can be aggregated to figure out what clever rejoinders you might want.
132
+
| `2` | similar to 0 (user first), but puts the name of the topic the chatbot ended in before either. You can see the flow of topics better with this view.
133
+
| `3` | similar to 2 (topic shown), but puts what the chatbot said before the user.
134
+
| `4` | puts the user and chatbot on separate lines, indenting the chatbots line. Easier to read.
135
+
| `5` | similar to 3, but indents the user instead of the chatbot.
136
+
| `6` | only lists the users inputs. This is good for creating a file that can recreate a user's experience, if you want to recreate it for debugging or regression.
137
+
| `7` | display rule responsible for output. Analogous to :why, it shows the rule tag, the sample input comment if there is one, the rule type and pattern, the input from the user and the output from the chatbot. If the rule doing the output was the target of a local `^reuse (same topic)`, then the data about the rule comes from the calling rule, not the output rule.
138
+
| `8` | puts the user and chatbot on separate lines, indenting the chatbots line and prefixes it with the topic generating the response. Easier to read and debug.
139
+
| `11`| puts the timestamp and user on first line and and chatbot on second line,indenting the chatbots line.
141
140
142
141
Normally trim displays everything. But with an optional 3rd argument `nooob`, you can omit out-of-bands data from output. E.g.,
143
142
```
144
143
:trim bob 6 nooob
145
144
```
146
-
You can separately choose to trim input and output oob using a numeric bits, where `1` is input and `2` is output (the same as `noob`) and
147
-
`3` is both
145
+
You can separately choose to trim input and output oob using a numeric bits, where `1` is input and `2` is output (the same as `noob`) and `3` is both
0 commit comments