mirror of
https://github.com/weewx/weewx.git
synced 2026-04-18 08:36:54 -04:00
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:v="urn:schemas-microsoft-com:vml">
<!-- $Revision$ -->
<!-- $Author$ -->
<!-- $Date$ -->
<head>
<meta content="en-us" http-equiv="Content-Language" />
<meta content="text/html; charset=utf-8" http-equiv="Content-Type" />
<title>The Weewx weather system</title>
<!-- CSS -->
<link href="weewx_docs.css" rel="stylesheet" />
<!-- JavaScript -->
<script src="samaxesjs.toc-1.3.2.min.js" type="text/javascript"></script>
<script type="text/javascript">samaxesJS.toc({exclude: 'h4, h5, h6', autoId: true, context: 'technical_content'});</script>
</head>
<body>
<h1 class="title">The <span class="code">weewx</span> weather system<br />
Version 1.10</h1>
<h1>Table of Contents</h1>
<p>For information on customizing <span class="code">weewx</span>, see the separate
document <a href="customizing.htm"><em>Customizing
<span class="code">weewx</span></em></a>.</p>
<p>For instructions on upgrading from various versions, see the separate document
<em><a href="upgrading.htm">Upgrading <span class="code">weewx</span></a></em>.</p>
<p>For instructions on porting to the SheevaPlug, see the separate document
<em><a href="sheeva.htm">Notes on porting <span class="code">weewx</span> to
the SheevaPlug</a></em>.</p>
<div id="technical_content">
<div id="toc">
</div>
<h1><a name="Copyright">Copyright</a></h1>
<p>(c) 2009, 2010, 2011 by Tom Keffer <<a href="mailto:tkeffer@gmail.com">tkeffer@gmail.com</a>></p>
<p>This program is free software: you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation, either version 3 of the License, or (at your option) any later version.
</p>
<p>This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
</p>
<p>You should have received a copy of the GNU General Public License along with
this program. If not, see <http://www.gnu.org/licenses/>
<a href="http://www.gnu.org/licenses/">http://www.gnu.org/licenses</a>.</p>
<h1><a name="About_weewx">About <span class="code">weewx</span></a></h1>
<p><span class="code">weewx</span> is a piece of software, written in
<a href="http://www.python.org">Python</a>, that interacts with your weather
station to produce plots, reports, and HTML pages. It can optionally upload
the reports to a remote Web server as well as publish to the
<a href="http://www.wunderground.com">WeatherUnderground</a> or
<a href="http://www.pwsweather.com/">PWSweather.com</a>. It uses modern software
concepts, making it simple, robust, and easy to extend. For an example station
see <a href="http://www.threefools.org/weewx">Hood River West</a>.</p>
<p>Key features:</p>
<ul>
<li>An easy to understand, simple, extensible micro-kernel architecture<em>;</em></li>
<li>Uploads to popular weather sites, such as Weather Underground and CWOP;</li>
<li>Uploads to your website using FTP;</li>
<li>Support for multiple <em>skins</em>;</li>
<li>Simple, but extensible templating system;</li>
<li>Support for multiple unit systems;</li>
<li>Calibration corrections;</li>
<li>Ability to extend <span class="code">weewx</span> with new services
and reports.</li>
</ul>
<p>I wrote <span class="code">weewx</span> over the winter of 2008-2009 for
two reasons: it was a wet and miserable winter here in Oregon with not much
else to do, so there was no good reason not to, and because I wanted a simple,
easy-to-understand server to run my Davis VantagePro2 weather station on a Linux
box. I had been using <a href="http://www.wviewweather.com/">wview</a>, which
is a high-performance and feature rich system authored by Mark Teel with lots
of users. Written in C, it's an efficient system that can run on underpowered
boxes. In exchange, it's huge (45,000+ lines of code), tightly integrated in
with its companion library, radlib (another 14,000+ lines), and very complex,
making it difficult to understand and reliably customize. I wanted something
more modern and much simpler.</p>
<p>Having made a career in C++ and Java, I was also interested in some more
modern languages, so I thought I'd try either Python or Ruby (although, truth
be told, the roots of Python are nearly as old as C++!). I ended up picking
Python because its libraries are more mature and there are many mores choices
for third party libraries.</p>
<p><span class="code">Weewx</span> weighs in at well under 5,000 lines of code.
It also has another 2,900 comment lines. Because it is pure Python, it requires
no makefiles, no builds, no special installs. It offers very powerful configuration
and templating options, as well as an internally extensible engine, making it
easy to customize. Its internal modular design and use of modern exception handling
make it very robust and difficult to crash. It is also architecturally very
simple and easy to understand. However, to be fair, at this point it supports
only the
<a href="http://www.davisnet.com/weather/products/vantagepro.asp">Davis VantagePro2</a>
weather station.</p>
<h1><a name="Downloading_weewx">Downloading <span class="code">weewx</span></a></h1>
<p><span class="code">Weewx</span> can be downloaded from its
SourceForge page:<a href="http://sourceforge.net/projects/weewx/">http://sourceforge.net/projects/weewx/</a>.
</p>
<h1><a name="Prerequisites">Prerequisites</a></h1>
<h2>Python</h2>
<p>Python V2.5 or V2.6 is required. The newer V3.0 distribution will not work.</p>
<h2>Required packages</h2>
<p>The following external packages are required to use <span class="code">weewx</span>.</p>
<ol>
<li><a href="http://www.sqlite.org/">sqlite3</a> (Version 3.5 or greater)
A SQL database written in C, which <span class="code">weewx</span> uses
to store data pulled from the weather station. Comes with Debian and many
other Linux distributions.</li>
<li><a href="http://pypi.python.org/pypi/pysqlite/">pysqlite</a> (Version
2.5 or greater) The Python interface to sqlite3.</li>
<li><a href="http://pypi.python.org/pypi/configobj/">configobj</a> (Version
4.5 or greater) Manages the configuration file <span class="code">weewx.conf</span>.</li>
<li><a href="http://pypi.python.org/pypi/pyserial">pyserial</a> (Version
2.3 or greater) Manages the serial connection to the weather station.</li>
<li><a href="http://www.cheetahtemplate.org">Cheetah</a> (Version 2.0 or
greater) The HTML templating engine.</li>
<li><a href="http://pypi.python.org/pypi/PIL">Python Imaging Library</a>
(Version 1.1.6 or greater) Also known as PIL, this is included in many Python
distributions.</li>
</ol>
<p>There are two general strategies for installing these prerequisites:</p>
<ol>
<li>Use operating system tools, such as <span class="code">apt-get</span>
(or its graphical equivalent Synaptic Package Manager) for Debian/Ubuntu
or <span class="code">yast</span> for SuSE; or</li>
<li>Use the Python tool <span class="code">
<a href="http://pypi.python.org/pypi/setuptools">easy_install</a></span>.</li>
</ol>
<p>Option #1 is easier, but if your Linux distribution does not come with such
tools, you may have to use <span class="code">easy_install</span>. Brief instructions
for both approaches are given below.</p>
<h3>Installation on Debian distributions (including Ubuntu) using
<span class="code">apt-get</span></h3>
<p>The instructions that follow are for using the Debian tool
<span class="code">apt-get</span>, but the same package names would be used
should you chose to use a graphical interface such as the Synaptic Package Manager.</p>
<p>To install the required packages:</p>
<p class="tty">sudo apt-get install sqlite3<br />
sudo apt-get install python-pysqlite2<br />
sudo apt-get install python-configobj<br />
sudo apt-get install python-serial<br />
sudo apt-get install python-cheetah<br />
sudo apt-get install python-imaging</p>
<p>Note that <span class="code">sqlite3</span> and the <span class="code">
python-imaging</span> (PIL) packages are preinstalled on many Debian
distributions so the corresponding commands may not do anything.</p>
<h3>Installation on SuSE using <span class="code">yast</span></h3>
<p>My SuSE 11.3 system came with some of the prerequisites installed, some available
through <span class="code">yast</span>, and some that required
<span class="code">easy_install</span>. </p>
<p>To start, some packages want to install themselves into
<span class="code">/usr/local/lib/python2.6/site-packages</span>, which may
or may not exist. If it does not, you may have to create it before
running <span class="code">easy_install</span>:</p>
<pre>sudo mkdir /usr/local/lib/python2.6</pre>
<pre>sudo mkdir /usr/local/lib/python2.6/site-packages</pre>
<p>Install the
gcc compiler:</p>
<pre>sudo yast -i gcc</pre>
<p>Then install <span class="code">easy_install</span>:</p>
<pre>sudo yast -i python-setuptools</pre>
<p>Install <span class="code">configobj</span>:</p>
<pre>sudo yast -i python-configobj</pre>
<p>Install PIL:</p>
<pre>sudo yast -i python-imaging</pre>
<p>Install <span class="code">pyserial</span>:</p>
<pre>sudo yast -i python-serial</pre>
<p>Install <span class="code">sqlite3</span>. My SuSE 11.3 system came with V3.6.4, which works just fine. However, if
you need to install:</p>
<p class="tty">sudo yast -i sqlite3</p>
<p>Install the Python development package (need for <span class="code">
pysqlite</span>):</p>
<pre>sudo yast -i python-devel</pre>
<p>Install <span class="code">pysqlite</span>:</p>
<pre>easy_install pysqlite</pre>
<p>Install <span class="code">Cheetah</span>:</p>
<pre>easy_install Cheetah</pre>
<h3>Installation using <span class="code">easy_install</span></h3>
<p>An alternative approach to installing the required packages is by using the
Python setup tool "<span class="code">easy_install</span>", part of the
<a href="http://pypi.python.org/pypi/setuptools">python-setuptools package</a>.
Refer to their instructions on how to install this tool.</p>
<p>Once <span class="code">easy_install</span> has been installed, installing
the rest of the packages is very easy.</p>
<p>My Ubuntu 8.10 system came with sqlite V3.5.9, which works just fine. If
you do not have sqlite3, refer to <a href="http://www.sqlite.org">the sqlite
webpage</a> for installation instructions.</p>
<p>While Version 2.3.X of <a href="http://pypi.python.org/pypi/pysqlite/">pysqlite</a>
is included with many versions of Python, the more recent 2.5.X or greater is
required in order to take advantage of transaction contexts. Hence, you may
have to install or upgrade. Because pysqlite builds a C library, you may have
to install the Python development environment first, if you have not already
done so. Generally, this means installing the gcc compiler. You may also have
to install the sqlite3 development environment as well.</p>
<p>With the development environment in place, you can easily build and install
pysqlite:</p>
<pre>easy_install pysqlite</pre>
<p>If your system already has a version of pysqlite installed, but it is not
a high enough version (<span class="code">easy_install</span> will tell you
the version you have), then you may have to force an upgrade:</p>
<pre>easy_install --upgrade pysqlite</pre>
<p>Note that at the time of this writing (24 Oct 2009), the hosting site for
pysqlite has changed to one on <a href="http://pysqlite.googlecode.com">googlecode</a>,
and <span class="code">easy_install</span> could not find it. You may have to
find and give the URL explicitly to <span class="code">easy_install</span> (adjust
version numbers as necessary):</p>
<pre>easy_install http://pysqlite.googlecode.com/files/pysqlite-2.5.5.tar.gz</pre>
<p>The other packages are straightforward. Install configobj:</p>
<pre>easy_install configobj</pre>
<p>Install pyserial:</p>
<pre><span class="code">easy_install pyserial</span></pre>
<p>Install Cheetah:</p>
<pre><span class="code">easy_install Cheetah</span></pre>
<p>My version of Python came with V1.1.6, which works great, but if you need
to install</p>
<pre>easy_install pil</pre>
<h2>System requirements</h2>
<p>I run <span class="code">weewx</span> on a 500MHz system with an AMD Geode
processor and 512 MB of memory. Configured this way, it consumes about
5% of the CPU and about 40MB of total memory.</p>
<h2>Weather station hardware requirements</h2>
<p>At this point, only the
<a href="http://www.davisnet.com/weather/products/vantagepro.asp">Davis VantagePro2</a>
is supported, and even then, only the "Revision B" version (firmware dated on
or after 22 April 2002). It would be very easy to port to a "Revision A" station
or even the original VantagePro, but I don't have access to the hardware to
test it.</p>
<h1><a name="Installing_weewx">Installing <span class="code">weewx</span></a></h1>
<p>Because <span class="code">weewx</span> is "Pure Python", that is it is 100%
Python with no "C" modules to compile, installing it is very easy. Furthermore,
it uses the standard Python
<a href="http://docs.python.org/install/index.html">distutils</a> install method,
which is very easy and flexible. Detailed instructions follow.</p>
<h2>Unpacking</h2>
<p>Start by unpack the tar ball (substitute your version for X.Y.Z) into any
convenient directory where you have write permission</p>
<p class="tty">tar xvf weewx-X.Y.Z</p>
<p>Then change directory into it:</p>
<p class="tty">cd weewx-X.Y.Z</p>
<h2>Choosing where to install <span class="code">weewx</span></h2>
<p>Next step is figuring out where you want to install <span class="code">weewx</span>.
If <span class="code"><em>$WEEWX_INSTALL</em></span> symbolizes the root location
of where the <span class="code">weewx</span> directory will be installed, then
the nominal directory layout is:</p>
<table align="center" style="width: 60%">
<tr>
<td><strong>Subdirectory</strong></td>
<td><strong>Contents</strong></td>
</tr>
<tr>
<td><span class="code"><em>$WEEWX_INSTALL</em>/bin</span></td>
<td>Python packages and scripts</td>
</tr>
<tr>
<td><span class="code"><em>$WEEWX_INSTALL</em>/weewx.conf</span></td>
<td>Configuration file</td>
</tr>
<tr>
<td><span class="code"><em>$WEEWX_INSTALL</em>/skins</span></td>
<td>Template skins and their configuration files</td>
</tr>
<tr>
<td><span class="code"><em>$WEEWX_INSTALL</em>/archive</span></td>
<td>sqlite3 archive and statistical databases</td>
</tr>
<tr>
<td><span class="code"><em>$WEEWX_INSTALL</em>/public_html</span></td>
<td>Generated HTML and .PNG images</td>
</tr>
</table>
<p>By default, the location for <span class="code"><em>WEEWX_INSTALL</em></span>
is <span class="code">/home/weewx</span>. However, it can be changed by editing
the file <span class="code">setup.cfg</span>. If you wish to install someplace
else, open up <span class="code">setup.cfg</span> and change the line</p>
<p class="tty">home = /home/weewx</p>
<p>to reflect your decision.</p>
<h2>Build and install</h2>
<p>Build the distribution</p>
<p class="tty">./setup.py build</p>
<p>(Because <span class="code">weewx</span> is pure Python this doesn't actually
build anything, but it does arrange files for the final installation)</p>
<p>Then install it. If you have write permission in the directory where
<span class="code">weewx</span> will go (<em>i.e.</em>, <span class="code">
<em>$WEEWX_INSTALL</em></span>), then type</p>
<p class="tty">./setup.py install</p>
<p>Otherwise, if you do not have write permission, you will have to use sudo:</p>
<p class="tty">sudo ./setup.py install</p>
<h3>Upgrading</h3>
<p>Before upgrading from a previous version of <span class="code">weewx</span>,
check the upgrade notes, found in file <a href="upgrading.htm">upgrading.htm</a>,
to see if there any specific actions you need to do. </p>
<p>You then generally follow the procedure above. </p>
<p><strong>In particular, before starting, be sure to set </strong>
<span class="code"><strong>home</strong></span><strong> in the file </strong>
<span class="code"><strong>setup.cfg</strong></span><strong>. </strong></p>
<p>The build and install process will do the following for you.</p>
<ul>
<li>Save your old '<span class="code">bin</span>' subdirectory as
<span class="code"><em>$WEEWX_INSTALL</em>/bin.YYYYMMDDHHMMSS</span> where
YYYYMMDDHHMMSS is a timestamp;</li>
<li>Install the new version in the '<span class="code">bin</span>'
subdirectory, while preserving any user extensions in the '<span class="code">bin/user</span>'
subdirectory;</li>
<li>Save a copy of your old <span class="code">weewx.conf</span> as
<span class="code"><em>$WEEWX_INSTALL</em>/weewx.conf.YYYYMMDDHHMMSS;</span></li>
<li>Merge any changes you've made to your old configuration file
<span class="code">weewx.conf</span> into the new configuration file, then
install the merged copy (this effectively causes changes you've made to
override the values in the shipped version of <span class="code">weewx.conf</span>);</li>
<li>Install a <span class="code">'skins'</span> subdirectory if one does
not already exist.</li>
</ul>
<h2>Permissions</h2>
<p>Strictly speaking it is not necessary to install or run
<span class="code">weewx</span> with root privileges. You only need read/write
access to the serial port for your hardware. For example, if your hardware has
a USB interface, on Ubuntu and SuSE:</p>
<p class="tty">sudo chmod a+rw /dev/ttyUSB0</p>
<p>Once done, if you edit <span class="code">setup.cfg</span> to install into
a directory where you have write permissions, you can install and run
<span class="code">weewx</span> without any root privileges at all. </p>
<h2>Final note on installation</h2>
<p>Because <span class="code">weewx</span> is pure Python, it does not actually
have to be "built" and "installed" at all! You can just simply run it out of
whatever directory you unpack it into. However, it must still be told where
to find the skin templates and databases, as well as where to put the generated
HTML files and plot images. This is done by setting <span class="code">WEEWX_ROOT</span>
in <span class="code">weewx.conf</span> to reflect where your data files can
be found (normally, the install process does this for you):</p>
<pre>> # Edit WEEWX_ROOT to reflect the location of your data hierarchy:
> vi ./weewx-X.Y.Z/weewx.conf
> # Now run weewxd.py out of the directory you unpacked weewx into:
> ./weewx-X.Y.Z/bin/weewxd.py ./weewx-X.Y.Z/weewx.conf</pre>
<h1><a name="Configuring_weewx">Configuring <span class="code">weewx</span></a></h1>
<p>This section covers configuring <span class="code">weewx</span>, in particular
the configuration files <span class="code">weewx.conf</span> and
<span class="code">skin.conf</span>.</p>
<p>In the following, <span class="code"><em>$WEEWX_ROOT</em></span> refers to
the <span class="code">weewx</span> data root directory, generally
<span class="code">/home/weewx</span>. The subdirectories <span class="code">
archive</span>, <span class="code">skins</span>, and <span class="code">public_html</span>
are expected to be found here. </p>
<h2>Moving from wview</h2>
<p>NOTE: This section is necessary only if you are moving from
<a href="http://www.wviewweather.com/">wview</a> to <span class="code">weewx</span>
and wish to transfer your old data over. If you are starting afresh, you do
not need to follow this section — the two main databases are created and populated
automatically by <span class="code">weewx</span>.</p>
<p>The main archive database can be just copied over because wview and
<span class="code">weewx</span> use identical schemas:</p>
<pre>mkdir <em>$WEEWX_ROOT</em>/archive
cp /usr/local/var/wview/archive/wview-archive.sdb <span class="code"><em>$WEEWX_ROOT</em></span>/archive/weewx.sdb</pre>
<p>The second database, the statistics, will then be automatically created and
backfilled on startup of weewx. If your existing database is large, backfilling
could take some time. On my modest 500 MHz
<a href="http://www.fit-pc.com/new/fit-pc-slim-specifications.html">fit-PC Slim</a>
with 512 MB of memory it took a little over 4 minutes for a year and a half
(25 MB) of data.</p>
<h2>Configuring your weather station</h2>
<p>The only two variables <span class="code">weewx</span> tries to manage on
the VantagePro are the time and the archive interval. </p>
<h3>Time</h3>
<p>The time on the VP is automatically synchronized with the
<span class="code">weewx</span> server nominally every four hours (changeable
by the user). You should run a <a href="http://www.ntp.org/">NTP</a> daemon
on your server to insure that it is synchronized with the correct time. Doing
so will greatly reduce errors, especially if you send data to services such
as the Weather Underground.</p>
<h3><a name="Archive_interval">Archive interval</a></h3>
<p>The archive interval is set in the main configuration file
<span class="code"><em>$WEEWX_ROOT</em>/weewx.conf. </span>Look for the entry
<span class="code">archive_interval</span> in the <span class="code">VantagePro</span>
section. Set it to the number of seconds. Valid entries are 60, 300, 600, 900,
1800, 3600, and 7200. However, if you are ftp'ing lots of files to a server,
setting it to 60 seconds may not give enough time to have them all uploaded
before the next archive record is due. If this is the case, you should pick
an archive interval of at least 300 seconds, or trim the number of files you
are using.</p>
<p>I have found that a five minute (300 seconds) archive interval works well
for the VantagePro. Because of the large amount of onboard memory it carries,
going to a larger interval really does not have any advantages.</p>
<p><em>Choose your archiving interval carefully! </em>Once chosen, it cannot
be changed without messing up your statistics (highs and lows will be OK, but
averages and rms wind speed will be wrong).</p>
<p>After setting the desired interval in the configuration file, run the
<span class="code">configure.py</span> script to set it on the VantagePro. If
it differs from the old archive interval, the main memory log of the VantagePro
will be cleared. </p>
<p class="tty"><em>$WEEWX_ROOT</em>/bin/configure.py --configure-VantagePro
$WEEWX_ROOT/weewx.conf</p>
<h2>Editing the configuration file <span class="code">weewx.conf</span></h2>
<p>Station specific information is set in the configuration file
<span class="code"><em>$WEEWX_ROOT</em>/weewx.conf</span>. There is another
configuration file <span class="code">skin.conf</span> for presentation-specific
options, which is described in a separate document <em>
<a href="customizing.htm">Customizing <span class="code">weewx</span></a>
</em>under section <em>
<a href="customizing.htm#Reference:_The_Standard_skin_configuration_file">Reference:
The standard skin configuration file</a></em>.</p>
<p>Most of the important options are up near the top of the file. They are all
documented in this section, although you can safely ignore most of them. The
truly important ones, the ones you are likely to have to customize for your
station, are shown in <span class="bold_n_blue"><strong>bold face and in blue</strong></span>.
</p>
<p>Default values are provided for many of them, meaning that if they are not
listed in the configuration file <em>at all</em>, <span class="code">weewx</span>
will pick sensible values. When the documentation below gives a "default value"
this is what it means. However, all options have been given values in the configuration
file that ships with <span class="code">weewx</span>, so you can see what they
look like. The value given in this shipped configuration file is not necessarily
the same as the "default value".</p>
<p>What follows is organized by the different sections of the configuration
file.</p>
<h3>General</h3>
<p>The options declared at the top are not actually part of any section. There
are two:</p>
<p class="config_option">debug</p>
<p>Set to 1 to have the program perform extra debug checks, as well as emit
extra information on the log file. Otherwise, set to 0. Default is 0 (no debug).</p>
<p class="config_option">socket_timeout</p>
<p>Set to how long to wait before declaring a socket time out. This is used
when FTP'ing data to a web server or sending data to the Weather Underground.
Twenty (20) seconds is reasonable. Default is 20.</p>
<h3 class="config_section">[Station]</h3>
<p>This section covers options relating to the entire weather station setup.
</p>
<p class="config_important">WEEWX_ROOT</p>
<p>Set to the root directory of the <span class="code">weewx</span> data file
hierarchy for this station, nominally '<span class="code">/home/weewx</span>'.
The weewx data subdirectories <span class="code">skins</span>,
<span class="code">archive</span>, and <span class="code">public_html</span>
are expected to be found here. This value will be set automatically by the setup
script <span class="code">setup.py</span> to reflect the choice you made in
the configuration file <span class="code">setup.cfg</span>. Required. No default.</p>
<p class="config_important">location</p>
<p>The station location should be a UTF-8 string that describes the geography
of where you weather station is located, such as '<span class="code">Hood River,
Oregon</span>'. Required. No default.</p>
<p class="config_important">latitude<br />
longitude</p>
<p>The lat/lon should be set in decimal degrees, negative for southern and eastern
hemispheres, respectively. Required. No default.</p>
<p class="config_important">altitude</p>
<p>Set to the altitude of the station and the unit used for the altitude. Required.
No default. Example:</p>
<pre>altitude = 700, foot</pre>
<p class="config_important">rain_year_start</p>
<p>If your area uses a rain year that starts on something other than the first
of January, you may want to set this variable. For example, set to 10 if your
rain year starts in October (as mine does). Default is 1.</p>
<p class="config_option">week_start</p>
<p>Start of the week. 0=Monday, 1= Tuesday, ... , 6 = Sunday. Default is 6 (Sunday)</p>
<p class="config_option">clock_check</p>
<p>How often to check the station's onboard clock for drift, in seconds. Default
is 14400 (every 4 hours)</p>
<p class="config_option">max_drift</p>
<p>The maximum amount of drift to tolerate, in seconds, in the onboard clock,
before resetting the clock. Default is 5.</p>
<p class="config_option">cache_loop_data</p>
<p>Set to 1 (one) to cache LOOP data, otherwise, set to zero. Most users will
not want to change this unless you have a specialized application. Default is
1 (<em>i.e.</em>, cache LOOP data).</p>
<p class="config_option">station_type</p>
<p>Set to the type of hardware you are using. For this version, only '<span class="code">VantagePro</span>'
is accepted. Required.</p>
<h3 class="config_section">[VantagePro]</h3>
<p>This section is for options relating to the VantagePro hardware.</p>
<p class="config_important">port</p>
<p>Set to the port name used by your station. Example, /<span class="code">dev/ttyUSB0</span>
is a common location for USB ports under Debian, <span class="code">/dev/ttyS0</span>
for serial ports. Required. No default.</p>
<p class="config_option">baudrate</p>
<p>Set to the baudrate of your station. The default is 19200.</p>
<p class="config_important">archive_interval</p>
<p>Set to the desired archive interval of your station, in seconds. This variable
is only used when setting up your station. Otherwise, this value is read directly
from the station. Required if you <a href="#Archive_interval">configure your
station</a>. No default.</p>
<p class="config_option">iss_id</p>
<p>Set to the ID number of your Integrated Sensor Suite (ISS). This is used
in the formula to calculate reception quality for wireless stations. The default
is 1.</p>
<p class="config_option">archive_delay</p>
<p>How long to wait in seconds after the top of an archiving interval before
fetching new data off the station. For example, if your archive interval is
5 minutes and archive_delay is set to 15, then the data will be fetched at 00:00:15,
00:05:15, 00:10:15, etc. This delay is to give the station a few seconds to
archive the data internally, and in case your server has any other tasks to
do at the top of the minute. Default is 15 seconds.</p>
<p class="config_option">timeout</p>
<p>How many seconds to wait for a response from the station before giving
up. Default is 5 seconds.</p>
<p class="config_option">wait_before_retry</p>
<p>How many seconds to wait before retrying again. Unless you have a good reason
to change it, this value should be left at the default, as it is long enough
for the station to offer new data, but not so long as to go into a new loop
packet (which arrive every 2 seconds). Default is 1.2 seconds.</p>
<p class="config_option">max_tries</p>
<p>How many times to try again before giving up. Default is 4.</p>
<p class="config_option">unit_system</p>
<p>What unit system is in use on your weather station hardware. Possible values
are '1' (U.S. Customary) or '2' (Metric). As far as I know, all Davis instruments
support only U.S.. In any case, U.S. is the only system supported by
<span class="code">weewx</span> at this time. Default is 1.</p>
<h3 class="config_section">[RESTful]</h3>
<p>This section is for configuring uploads to simple
<a href="http://en.wikipedia.org/wiki/Representational_State_Transfer">RESTful</a>
servers such as the <a href="http://www.wunderground.com/">Weather Underground</a>,
<a href="http://www.pwsweather.com/">PWSweather.com</a>, or
<a href="http://www.wxqa.com/">CWOP</a>.</p>
<h4 class="config_section">[[Wunderground]]</h4>
<p><span class="code">Weewx </span>can send your current data to the Weather
Underground. If you do not wish to do this, comment out the two options below.</p>
<p class="config_important">station</p>
<p>Set to your Weather Underground station ID (e.g., <span class="code">KORHOODR3</span>).
Required.</p>
<p class="config_important">password</p>
<p>Set to your Weather Underground password. Required.</p>
<h4 class="config_section">[[PWSweather]]</h4>
<p><span class="code">Weewx </span>can send your current data to the PWSweather.com
service. If you do not wish to do this, comment out the two options below.</p>
<p class="config_important">station</p>
<p>Set to your PWSweather station ID. Required.</p>
<p class="config_important">password</p>
<p>Set to your PWSweather password. Required.</p>
<h4 class="config_section">[[CWOP]]</h4>
<p><span class="code">Weewx</span> can send your data to the Citizen Weather
Observer Program. If you do not wish to do this, comment out option "<span class="code">station</span>"
below.</p>
<p class="config_important">station</p>
<p>Set to your CWOP station ID (e.g., <span class="code">CW1234</span>). Required.</p>
<p class="config_important">passcode</p>
<p>This is used for APRS (amateur radio) stations only. Set to the passcode
given to you by the CWOP operators. Otherwise, leave this option commented out
[Required for APRS stations; ignored for others]</p>
<p class="config_option">server</p>
<p>A comma separated list of servers:ports to try. Unless you have a good reason
to change this, use the servers listed. [Required.]</p>
<p class="config_option">interval</p>
<p>The interval in seconds between posts. Because CWOP is heavily used, the
operators discourage very frequent posts. Every 5 minutes (300 seconds) is fine,
but they prefer every 10 minutes (600 s) or even longer. Setting this value
to zero will cause every archive record to be posted. [Optional. Default is
zero.]</p>
<p class="config_option">stale</p>
<p>How old a record can be before it will not be used for a catch up. CWOP does
not use the timestamp on a posted record. Instead, they use the wall clock time
that it came in. This means that if your station is off the air for a long period
of time, then weewx attempts a catch up, old data could be interpreted as the
current conditions. [Optional. Default is 1800 seconds.]</p>
<h3 class="config_section">[Archive]</h3>
<p>This section is for configuring the sqlite3 database in which the station
archive data is stored.</p>
<p class="config_option">archive_file</p>
<p>The path, relative to the <span class="code">WEEWX_ROOT </span>directory,
to the database. Required</p>
<p class="config_option">unit_system</p>
<p>What unit system to use inside the database. Required. The only one supported
right now is '1', the U.S. Customary system</p>
<h3 class="config_section"><a name="[Stats]">[Stats]</a></h3>
<p>This section is for configuring the sqlite3 database in which the station
statistics are stored.</p>
<p class="config_option">stats_file</p>
<p>The path, relative to the <span class="code">WEEWX_ROOT </span>directory
to the statistical database. Required.</p>
<p class="config_option"><a name="stats_types">stats_types</a></p>
<p>The list of types for which statistics will be kept. Types not listed will
not be available for generating HTML pages. Optional. The default is all types,
resulting in a possibly much bigger than necessary stats database (do you really
have four different soil moisture sensors?) The list that ships with the configuration
file will work for most stations and probably will not have to be modified.</p>
<h3 class="config_section"><a name="[Reports]">[Reports]</a></h3>
<p>This section controls which reports are to be generated. While it can be
highly customized for your individual situation, this documentation describes
the section as shipped in the standard distribution.</p>
<p>Each report is represented by a sub-section, marked with double brackets
(<em>e.g.</em>, <span class="code">[[MyReport]]</span>). Any options for the
report should be placed under it. The standard report service will go through
the sections, running each report in order. Hence, for the stock distribution,
report <span class="code">[[StandardReport]]</span> will be run first, then
report <span class="code">[[FTP]]</span> (which actually optionally uploads
the results to a remote web server). Details for how to customize reports are
in the section <em>
<a href="customizing.htm#Opportunities_for_customizing_reports">Opportunities
for customizing reports</a></em>, in the document <em>
<a href="customizing.htm">Customizing <span class="code">weewx</span></a></em>.</p>
<p class="config_option">SKIN_ROOT</p>
<p>The directory relative to <span class="code"><em>$WEEWX_ROOT</em></span>
where the skins live. Default is <span class="code">skins</span>.</p>
<p class="config_option">HTML_ROOT</p>
<p>The target directory for the generated files, relative to
<span class="code"><em>$WEEWX_ROOT</em></span>. Generated files and images will
be put here. Default is <span class="code">public_html</span>.</p>
<h4 class="config_section">[[StandardReport]]</h4>
<p>This is the standard report that will be run on every archiving interval.
It uses the skin "<span class="code">Standard</span>", which generates four
HTML pages ("day", "week", "month", and "year" observations), plot graphs for
same, an RSS feed, and NOAA monthly and yearly reports. Unless changed otherwise,
it uses US Customary Units and puts the results in <span class="code">public_html</span>
and subdirectory <span class="code">public_html/NOAA</span>.</p>
<h4 class="config_section">[[FTP]]</h4>
<p>While this "report" doesn't actually generate anything, it uses the report
machinery to upload files from directory <span class="code"><em>$HTML_ROOT</em></span>
to a remote webserver. It does an incremental update, that is, it only FTPs
any files that have changed, saving the outgoing bandwidth of your Internet
connection.</p>
<p>If you do not use such a server, comment out the first four options below.</p>
<p class="config_important">user</p>
<p>Set to the username you use for your FTP connection to your web server. Required.
No default.</p>
<p class="config_important">password</p>
<p>Set to the password you use for your FTP connection to your web server. Required.
No default.</p>
<p class="config_important">server</p>
<p>Set to the name of your web server (e.g.,
<a href="http://www.threefools.org">www.threefools.org</a>, in my case). Required.
No default</p>
<p class="config_important">path</p>
<p>Set to the path where the weather data will be stored on your webserver (e.g.,
'<span class="code">/weather</span>'). NB: some FTP servers require a leading
slash ('<span class="code">/</span>'), some don't. Required. No default.</p>
<p class="config_important">passive</p>
<p>Set to 1 if you wish to use the more modern, FTP passive mode, 0 if you wish
to use active mode. Passive mode generally works better through firewalls, but
not all FTP servers do a good job of supporting it. See
<a href="http://slacksite.com/other/ftp.html">Active FTP vs. Passive FTP, a
Definitive Explanation</a> for a good explanation of the difference. Default
is 1 (passive mode).</p>
<p class="config_option">max_tries</p>
<p><span class="code">Weewx</span> will try up to this many times to FTP a file
up to your server before giving up. Default is 3.</p>
<h3 class="config_section">[<a name="Calibrate">Calibrate</a>]</h3>
<p>One of the services that <span class="code">weewx</span> can run is a <em>
calibration service</em> called <span class="code">StdCalibrate</span>. This
service offers an opportunity to correct for any calibration errors in your
instruments. It is very general and flexible.</p>
<p>If you do not wish to apply any calibrations, you can leave it out of
<span class="code"><a href="#service_list">service_list</a></span>, the list
of services to be run, and it will not be loaded or run.</p>
<h4 class="config_section">[[Corrections]]</h4>
<p>In this section you list all <em>correction expressions</em>. For example,
say that you know your outside thermometer reads high by 0.2°F. You could add
the expression:</p>
<pre>outTemp = outTemp - 0.2</pre>
<p>Perhaps you need a linear correction around a reference temperature of 68°F:</p>
<pre>outTemp = outTemp + (outTemp-68) * 0.02</pre>
<p>It is even possible to do corrections involving more than one variable. Suppose
you have a temperature sensitive barometer:</p>
<pre>barometer = barometer + (outTemp-32) * 0.0091</pre>
<p>All correction expressions are run in the order given.</p>
<p>The units should be <em>in the native units of the instrument</em>, US Customary
in the case of the VantagePro.</p>
<p>Both LOOP data and archive data will be corrected. It is important that the
calibration service, <span class="code">StdCalibrate</span>, be run <em>before</em>
the archiving service <span class="code">StdArchive</span>, so that it is the
corrected data that is stored.</p>
<p>If all you require is a simple correction offset, this can also be done in
the Davis VantagePro hardware. See your manual for instructions.</p>
<h3 class="config_section">[<a name="QC">QC</a>]</h3>
<p>Another service that weewx can run is a <em>quality control service</em>
called <span class="code">StdQC</span>. This service is very simple and only
checks that values are within a minimum and maximum range.</p>
<p>If you do not wish to use this service, you can leave it out of
<span class="code"><a href="#service_list">service_list</a></span>, the list
of services to be run, and it will not be loaded or run.</p>
<h4 class="config_section">[[MinMax]]</h4>
<p>In this section you list the observation types you wish to have checked,
along with their minimum and maximum values. The units should be in the native
units of the weather station hardware (<em>e.g.</em>, US Customary in the case
of the Davis VantagePro2). For example,</p>
<p class="tty">[[MinMax]]<br />
outTemp = -40, 120<br />
barometer = 28, 32.5<br />
outHumidity = 0, 100</p>
<p>In this example, if a temperature should fall outside of the inclusive range
-40 °F through 120 °F, then it will be set to the null value,
<span class="code">None</span> and ignored. In a similar manner, the acceptable
values for barometric pressure would be 28 through 32.5 inHg, for humidity 0
through 100%.</p>
<p>Both LOOP and archive data will be checked. It is important that the quality
control service be run <em>after</em> the calibration service
<span class="code">StdCalibrate</span> and <em>before</em> the archiving service
<span class="code">StdArchive</span>, so that it is the calibrated data that
is checked and then stored.</p>
<p>Knowing the details of how your hardware encodes data helps to minimize the
number of observations that need to be checked. For example, the VP2 devotes
only one unsigned byte to storing wind speed, and even then
<span class="code">0xff</span> is devoted to a bad value, so the only possible
values that could appear are 0 through 126 mph, a reasonable range. So, there
is no real point in checking wind speed for these instruments.</p>
<h3 class="config_section">[Engines]</h3>
<p>This section is used to configure the internal service engine in weewx. It
is for advanced customization. Details on how to do this is found in the section
<em>
<a href="file://RAVEN/tkeffer/workspace/weewx/docs/customizing.htm#Customizing_the_weewx_service_engine">
Customizing the weewx service engine</a> </em>in the document
<a href="customizing.htm"><em>Customizing </em><span class="code"><em>weewx</em></span></a>.</p>
<h4 class="config_section">[[WxEngine]]</h4>
<p>This section is for options used by the service engine.</p>
<p class="config_option"><a name="service_list">service_list</a></p>
<p>This option is the list of <em>services</em> that are to be run by the service
engine. After each event (such as the arrival of LOOP data, etc.), they will
be run in the given order. The standard list of services run by weewx is:</p>
<p class="tty">service_list = weewx.wxengine.StdCalibrate, weewx.wxengine,StdQC,
weewx.wxengine.StdArchive, weewx.wxengine.StdTimeSynch, weewx.wxengine.StdPrint,
weewx.wxengine.StdRESTful, weewx.wxengine.StdReportService</p>
<p>You can leave some of these services out if you do not need them. However,
this will only make a slight difference in execution speed.</p>
<h1><a name="Running_weewx">Running <span class="code">weewx</span></a></h1>
<p><span class="code">Weewx</span> can be run either from the command line (useful
for diagnostic purposes because it will print out a summary of every LOOP data),
or as a daemon. When first trying <span class="code">weewx</span>, it's probably
best to run it from the command line because you will be able to see command
line diagnostics, as well as log messages.</p>
<h2>Running from the command line</h2>
<p><span class="code">Weewx</span> can easily be run from the command line.
Start by making sure you have appropriate permissions to the serial port your
weather station uses. For example, if you are using a plain old serial port:</p>
<p class="tty">sudo chmod 666 /dev/ttyS0</p>
<p>Then run the main loop program, <span class="code">weewxd.py</span>, giving
the configuration file as its only parameter:</p>
<p class="tty"><em>$WEEWX_ROOT</em>/bin/weewxd.py <em>$WEEWX_ROOT</em>/weewx.conf</p>
<p>It should start by downloading any archive data from your weather station
into the database <span class="code"><em>$WEEWX_ROOT</em>/archive/weewx.sdb</span>.
As the Davis VantagePro can store a couple thousand archive records internally,
this could take a minute or two. I've found this process particularly slow on
SuSE for some reason.</p>
<p><span class="code">Weewx</span> will then start monitoring LOOP data, printing
a short version of the received data on standard output, about once every two
seconds.</p>
<p>You can tell a running instance of <span class="code">weewx</span> to reread
its configuration file by sending it the <span class="code">HUP</span> signal.
First run <span class="code">ps</span> to find out the Process ID (PID) number
of the instance, then send it the <span class="code">HUP</span> signal:</p>
<p class="tty">ps -a # Note
the PID of the weewxd.py process</p>
<p class="tty">kill -HUP <em>pid</em> # Send it a HUP signal</p>
<h2><a name="Running_as_a_daemon">Running as a daemon</a></h2>
<p>For unattended operations it is best to have <span class="code">weewx</span>
run as a daemon, started automatically when the server is rebooted. Start by
selecting the appropriate run script. They can be found under
<span class="code"><em>$WEEWX_ROOT</em>/start_script</span>. </p>
<table align="center" style="width: 70%">
<tr>
<td style="width: 25%">SuSE:</td>
<td class="code"><em>$WEEWX_ROOT</em>/start_script/SuSE/weewx</td>
</tr>
<tr>
<td style="width: 25%">Debian/Ubuntu:</td>
<td class="code"><em>$WEEWX_ROOT</em>/start_script/Debian/weewx</td>
</tr>
</table>
<p>Check the chosen script to make sure the variable <span class="code">WEEWX_ROOT</span>
inside has been set to the proper root directory for your <span class="code">
weewx</span> installation (it should have been set to the correct value automatically
by the install process, but it's worth checking). </p>
<p>Copy it to the proper location for your system:</p>
<table align="center" style="width: 70%">
<tr>
<td style="width: 25%">SuSE:</td>
<td class="code">cp <em>$WEEWX_ROOT</em>/start_script/SuSE/weewx /etc/init.d</td>
</tr>
<tr>
<td>Debian/Ubuntu:</td>
<td class="code">cp <em>$WEEWX_ROOT</em>/start_script/Debian/weewx /etc/init.d</td>
</tr>
</table>
<p>Make sure the script is executable:</p>
<table align="center" style="width: 70%">
<tr>
<td style="width: 25%">SuSE:</td>
<td class="code">chmod +x /etc/init.d/weewx</td>
</tr>
<tr>
<td style="width: 25%">Debian/Ubuntu:</td>
<td class="code">chmod +x /etc/init.d/weewx</td>
</tr>
</table>
<p>Create symbolic links in the run level directories:</p>
<table align="center" style="width: 70%">
<tr>
<td style="width: 25%">SuSE:</td>
<td class="code">/usr/lib/lsb/install_initd /etc/init.d/weewx</td>
</tr>
<tr>
<td style="width: 25%">Debian/Ubuntu:</td>
<td class="code">update-rc.d weewx defaults 98</td>
</tr>
</table>
<p><span class="code">Weewx</span> will now start automatically whenever your
system is booted. You can also manually start, stop, and restart the
<span class="code">weewx</span> daemon:</p>
<p class="tty">/etc/init.d/weewx start</p>
<p class="tty">/etc/init.d/weewx stop</p>
<p class="tty">/etc/init.d/weewx restart</p>
<p>By default, the scripts are designed to have <span class="code">weewx</span>
run at run levels 2, 3, 4 and 5. Incidentally, a nice tool for setting run levels
with Debian (Ubuntu) systems is
<a href="http://sysv-rc-conf.sourceforge.net/">sysv-rc-conf</a>. It uses a curses
interface to allow you to change easily which run level any of your daemons
runs at. There is a similar tool on SuSE. From the start menu run the YAST Control
Center, then look for Systems Services (Runlevel). Pick "Expert" mode to see
the run levels.</p>
<p>You can also tell <span class="code">weewx</span> to reread its configuration
file without stopping by using the 'reload' option:</p>
<p class="tty">/etc/init.d/weewx reload</p>
<h1><a name="Compatibility_with_wview">Compatibility with <span class="code">
wview</span></a></h1>
<p>The sqlite3 archive database used by <span class="code">weewx</span> (nominally,
<span class="code">weewx.sdb</span>) is completely compatible with the database
used by <a href="http://www.wviewweather.com">wview</a> (usually called
<span class="code">wview-archive.sdb</span>), at least as of wview Version 5.2.X.
The schema and its semantics is identical. However, the statistical file
<span class="code">stats.sdb</span> is different, and must be rebuilt. This
will be done automatically on startup by <span class="code">weewx</span>.</p>
<h1><a name="Monitoring_weewx">Monitoring <span class="code">weewx</span></a></h1>
<p><span class="code">Weewx</span> logs many events to the system log. On Debian
systems, this is <span class="code">/var/log/syslog</span>, on SuSE,
<span class="code">/var/log/messages</span>. Your system may use yet another
place. When troubleshooting the system, be sure to check it!</p>
<p>Setting the option <span class="code">debug </span>in <span class="code">
weewx.conf</span> to <span class="code">1</span> (one) will generate many more
checks and output and can be useful for debugging.</p>
<h1> <a name="Troubleshooting">Troubleshooting</a></h1>
<p>If you get stuck, be sure to </p>
<ul>
<li>Set option <span class="code">debug </span>in <span class="code">weewx.conf</span>
to <span class="code">1</span> (one)! It will put lots more information
in the log file, which can be very useful for troubleshooting and debugging!</li>
<li><a href="#Monitoring_weewx">Look at the log file</a>. I am always happy
to take questions, but the first thing I will ask you is: "Did you look
at the log file?"</li>
<li>Run from the command line. Generally, weewx will catch and log any unrecoverable
exceptions. But if you are getting strange results, it is worth running
from the command line and looking for any clues.</li>
</ul>
<h2> Execution</h2>
<h3><span class="code">configobj</span> errors</h3>
<p>These are errors in the configuration file. Two are very common. Incidentally,
these errors are far easier to diagnose when <span class="code">weewx</span>
is run from the command line.</p>
<h4><span class="code">configobj.DuplicateError</span> exception</h4>
<p>This error is caused by using an identifier more than once in the configuration
file. For example, you may have inadvertently listed your FTP server twice:</p>
<pre>[Reports]
[[FTP]]
... (details elided)
user = fred
server = ftp.myhost.com
password = mypassword
server = ftp.myhost.com # OOPS! Listed it twice!
path = /weather
... </pre>
<p>Generally, if you encounter this error, the log file will give you the line
number it happened in:</p>
<pre>Apr 24 12:09:15 raven weewx[11480]: wxengine: Error while parsing configuration file /home/weewx/weewx.conf
Apr 24 12:09:15 raven weewx[11480]: wxengine: Unable to initialize main loop:
Apr 24 12:09:15 raven weewx[11480]: **** Duplicate keyword name at line 254.
Apr 24 12:09:15 raven weewx[11480]: **** Exiting. </pre>
<h4><span class="code">configobj.NestingError</span> exception</h4>
<p>This is a very similar error, and is caused by a misformed section nesting.
For example:</p>
<pre>[Reports]
[[FTP]]]
... (details elided)</pre>
<p>Note the extra closing bracket on the subsection <span class="code">FTP</span>.</p>
<h2>Hardware</h2>
<h3>Tips on a reliable system</h3>
<p>If you are having problems keeping your weather station up for long periods
of time, here are some tips, in decreasing order of importance:</p>
<ul>
<li>Run on dedicated hardware. If you are using the server for other tasks,
particularly as your desktop machine, you will have reliability problems.
If you are using it as a print or network server, you will probably be OK.</li>
<li>Run headless. Modern graphical systems are extremely complex. As new
features are added, test suites don't always catch up. Your system will
be much more reliable if you run it without a windowing system (X Windows,
in the case of Linux).</li>
<li>Use an Uninterruptible Power Supply (UPS). The vast majority of power
glitches are very short lived — just a second or two — so you do not need
a big one. The 425 VA unit I use to protect my fit-PC cost me $55 at Best
Buy.</li>
<li>Use a VantagePro console with a serial connection, not a USB connection.
See the next section for details.</li>
<li>If you do use a USB connection, put a ferrite coil on each end of the
connection.</li>
</ul>
<h3>cp2101 converter problems</h3>
<p>The USB converter used in the Davis VantagePro is known to have some "noise"
problems. The symptom is that the Linux kernel will disconnect from your old
USB port claiming "EMI noise", and reconnect to a new and different port, where
<span class="code">weewx</span> can't find it. Here's a typical log output:</p>
<pre>Nov 29 10:40:21 hummingbird kernel: [6661624.786792] hub 2-0:1.0: port 3 disabled by hub (EMI?), re-enabling...
Nov 29 10:40:21 hummingbird kernel: [6661624.786871] usb 2-3: USB disconnect, address 2
Nov 29 10:40:21 hummingbird kernel: [6661624.795778] cp2101 2-3:1.0: device disconnected
Nov 29 10:40:21 hummingbird weewx[25808]: VantagePro: Max retries exceeded while getting LOOP packets</pre>
<pre> </pre>
<pre>... (messages elided)</pre>
<pre> </pre>
<pre>Nov 29 10:40:22 hummingbird kernel: [6661625.352340] cp2101 2-3:1.0: cp2101 converter detected
Nov 29 10:40:22 hummingbird kernel: [6661625.528107] usb 2-3: reset full speed USB device using ohci_hcd and address 3
Nov 29 10:40:22 hummingbird kernel: [6661625.735497] usb 2-3: cp2101 converter now attached to ttyUSB1</pre>
<p>In this example, the VantagePro was connected to <span class="code">/dev/ttyUSB0</span>,
but then reconnected to <span class="code">/dev/ttyUSB1</span>. </p>
<p>If you put ferrite coils on the USB connection, you will eliminate 90% of
this problem. I did this about 6 months ago, and have not had a problem since.
</p>
<p>However, there is one final step you can take that will reallly harden up
your system: install a <span class="code">udev</span> script that will create
a symbolic link to the VantagePro USB port, whatever it might be. With this
approach, if the port jumps from <span class="code">ttyUSB0</span> to
<span class="code">ttyUSB1</span>, the symbolic link will move as well. You
just specify port <span class="code">/dev/vpro</span> and be done with it. Here's
how:</p>
<h4>Installing a udev script</h4>
<p>I have installed a file <span class="code">/etc/udev/rules.d/66-vpro.rules</span>
on my fit-PC that looks like this:</p>
<pre># Automount the VantagePro2 to port /dev/vpro.</pre>
<pre># Install in /etc/udev/rules.d/66-vpro.rules</pre>
<pre>#</pre>
<pre>ACTION=="add", ATTRS{interface}=="CP2102 USB to UART Bridge Controller", SYMLINK+="vpro"</pre>
<p>What this rule says is that when the USB port is plugged in (action
<span class="code">add</span>), and it has an attribute with name
<span class="code">interface</span> that is equal to "<span class="code">CP2102
USB to UART Bridge Controller</span>", then add a symbolic link for its physical
port to <span class="code">/dev/vpro</span>. </p>
<p>Your controller may have a different identifier!! I can recommend this article,
"<a href="http://www.reactivated.net/writing_udev_rules.html"><em>Writing udev
rules</em></a>," for how to find and write an appropriate
<span class="code">udev</span> rule for your controller. (Note, however, that
this article uses the old <span class="code">udevinfo</span> command, rather
than the newer <span class="code">udevadm</span> command.) In particular, run
the command</p>
<pre># udevadm info --attribute-walk --path $(udevadm info --query=path --name=/dev/ttyUSB0) </pre>
<p>where<span class="code"> /dev/ttyUSB0</span> is the port (substitute your
real USB port) the VP2 is attached to. It will print out various identifiers
that can be useful in identifying your VP2 to <span class="code">udev</span>.
While the example script above used a rule that matched attribute
<span class="code">interface</span>, others are possible. For example, you might
chose to match the attribute <span class="code">product</span>.</p>
<p>Once you've installed your <span class="code">udev</span> rule, you can then
set <span class="code">port=/dev/vpro</span> in <span class="code">weewx.conf</span>,
confident that it will always point to your VantagePro2, no matter which USB
port it is actually attached to!</p>
<p>I have tested this system many times. You can yank the USB port out of the
machine and then plug it back in while also pulling out the network connection
in the middle of an FTP upload: weewx will recover.</p>
<p>Or, at least, it should!</p>
<h2> Templates</h2>
<h3><span class="code">Cheetah.NameMapper.NotFound</span> errors</h3>
<p>If you get errors of the sort:</p>
<pre>Apr 12 05:12:32 raven reportengine[3074]: filegenerator: Caught exception "<class 'NameMapper.NotFound'>"
Apr 12 05:12:32 raven reportengine[3074]: **** Message: "cannot find 'fubar' in template /home/weewx/skins/Standard/index.html.tmpl"
Apr 12 05:12:32 raven reportengine[3074]: **** Ignoring template and continuing.</pre>
<p>you have a tag in your template that <span class="code">weewx</span> does
not recognize (in this example, it's the tag <span class="code">$fubar</span>
in the template <span class="code">/home/weewx/skins/Standard/index.html.tmpl</span>.</p>
<h1> A<a name="Architectural_notes">rchitectural notes</a></h1>
<p>This section is not needed to get started. </p>
<h2>Goals</h2>
<p>The primary goals of <span class="code">weewx </span>are:</p>
<ul>
<li>Architectural simplicity. No semaphores, no named pipes, no inter-process
communications, no complex multi-threading to manage. </li>
<li>"Fast enough." In any design decision, architectural simplicity and
elegance trumps speed. </li>
<li>One code base. The same code base should be used for all platforms,
all weather stations, all reports, and any combination of features. Ample
configuration and customization options should be provided so the user doesn't
feel tempted to start hacking code. At worse, the user may have to subclass,
which is much easier to port to newer versions of the code base, than customizing
the base code.</li>
<li>Minimal reliance on external packages, so the user doesn't have to go
chase them down all over the Web before getting started.</li>
<li>Support only the Davis VantagePro2 initially (that's what I have), but
make no architectural decisions that lock out other stations.</li>
<li>As "pythonic" as I know how to make it. I'm a beginner Python programmer
with two decades of experience in C++. I tried hard to not make the code
base look like it was written by a C++ programmer who stumbled across a
Python manual!</li>
</ul>
<h2>Strategies</h2>
<p>To meet these goals, the following strategies were used:</p>
<ul>
<li>A "micro-kernel" design. The actual internal engine does very little.
It's main job is to load and run <em>services</em> at runtime, making it
easy for users to add or subtract features.</li>
<li>A largely stateless design style. For example, many of the processing
routines read their own data from the database, rather than caching it and
sharing with other processing routines. While this means the same data may
be read multiple times, it also means the only point of possible cache incoherence
is through the database, where transactions are easily controlled. This
greatly reduces the chances of corrupting the data, making it much easier
to understand and modify the code base.</li>
<li>Isolate the data collection and archiving code in a single thread that
is simple enough that it is unlikely to crash. The report processing is
where most mistakes are likely to happen, so isolate that in a separate
thread. If it crashes, it will not affect the main data thread.</li>
<li>A powerful configuration parser,
<a href="http://www.voidspace.org.uk/python/configobj.html">ConfigObj</a>,
by Michael Foord and Nicola Larosa, was chosen to read the configuration
file. This allows many options that might otherwise have to go in the code
to go instead in a configuration file.</li>
<li>A powerful templating engine,
<a href="http://www.cheetahtemplate.org/">Cheetah</a>, was used. This allows
many variables that I may not have thought of to be accessed from within
the HTML templates, without starting to modify the code.</li>
<li>Pure Python. The code base is 100% Python — no underlying C libraries
need be built to install <span class="code">weewx</span>. This also means
no Makefiles are needed.</li>
</ul>
<p>While <span class="code">weewx </span>is nowhere near as fast at generating
images and HTML as its predecessor, <span class="code">wview </span>(this is
partially because it uses fancier fonts and a much more powerful templating
engine), it is 'fast enough' for all platforms but the slowest. I run it regularly
on a 500 MHz machine where generating the 9 images used in the "Current Conditions"
page takes just under 2 seconds (compared with <span class="code">wview</span>'s
0.4 seconds). </p>
<p>Unfortunately, the architectural goal of one code base is likely to be broken
with the arrival of Python V3.X. It has so many changes that are not backwards
compatible with V2.X, that a separate code base will most likely be needed.
My intention is to stick with the V2.5 and V2.6 versions until V3.X is so widespread
it cannot be ignored, then make a permanent switch. I doubt this will affect
the average <span class="code">weewx</span> user. </p>
<p>All writes to the databases are protected by transactions. You can kill the
program at any time (either Control-C if run from the command line or "<span class="code">/etc/init.d/weewx
stop</span>" if a daemon) without fear of corrupting the databases.</p>
<p>The code makes ample use of exceptions to insure graceful recovery from problems
such as network outages. It also monitors socket and console timeouts, restarting
whatever it was working on several times before giving up. In the case of an
unrecoverable console error (such as the console not responding at all), the
program waits 60 seconds then restarts the program from the top.</p>
<p>Any "hard" exceptions, that is those that do not involve network and console
timeouts and are most likely due to a logic error, are logged, reraised, and
ultimately cause thread termination. If this happens in the main thread (not
likely due to its simplicity), then this causes program termination. If it happens
in the report processing thread (much more likely), then only the generation
of reports will be affected — the main thread will continue downloading data
off the instrument and putting them in the database. You can fix the problem
at your leisure, without worrying about losing any data.</p>
<h2>Terminology</h2>
<p>This is a glossary of terminology used throughout the code.</p>
<table style="width: 100%">
<tr>
<td>packet</td>
<td>Something obtained off the weather station. Frequently uses a complex
internal encoding, so it requires some processing to be useful.</td>
</tr>
<tr>
<td>record</td>
<td>Something obtained off the SQL database. </td>
</tr>
<tr>
<td>archive packet</td>
<td>A packet obtained off the store on the weather station. For example,
with a Davis VantagePro, it's obtained using their
<span class="code">DMPAFT</span> command. </td>
</tr>
<tr>
<td>loop packet</td>
<td>A packet with the current observations. For example, with a Davis
VantagePro, it's obtained using their <span class="code">LOOP</span>
command. </td>
</tr>
<tr>
<td>archive record</td>
<td>A record obtained off the SQL database</td>
</tr>
<tr>
<td>tuple-time</td>
<td>An instance of the Python object <span class="code">time.struct_time</span>.
This is a 9-wise tuple that represent a time. It could be in either
local time or UTC, though usually the former. See module
<span class="code">time</span> for more information. They are useful
because they are a little closer in format to what the Davis VantagePro
uses, although they still require a bit of processing. Variables carrying
tuple time usually have a suffix '<span class="code">_tt</span>'.</td>
</tr>
<tr>
<td>epoch time</td>
<td>Sometimes referred to as "unix time," or "unix epoch time." The
number of seconds since the epoch, which is 1 Jan 1970 00:00:00 UTC.
Hence, it always represents UTC (well.... after adding a few leap seconds.
But, close enough). This is the time used on the sqlite archive and
appears as type '<span class="code">dateTime</span>' in the SQL schema,
perhaps an unfortunate name because of the similarity to the Python
type '<span class="code">datetime</span>'. Very easy to manipulate, but it's an opaque big number.
</td>
</tr>
<tr>
<td>time stamp</td>
<td>A variable in unix epoch time. Always in UTC. Variables carrying
a time stamp usually have a suffix '<span class="code">_ts</span>'.</td>
</tr>
<tr>
<td>datetime</td>
<td>An instance of the Python object <span class="code">datetime.datetime</span>.
Variables of type datetime usually have a suffix '<span class="code">_dt</span>'.</td>
</tr>
<tr>
<td>SQL type</td>
<td>A type that appears in the SQL database. This usually looks something
like '<span class="code">outTemp</span>', '<span class="code">barometer</span>',
'<span class="code">extraTemp1</span>', and so on.</td>
</tr>
<tr>
<td>observation type</td>
<td>A type that can be used in the presentations. This is generally
all of the SQL types, plus calculated data (such as
<span class="code">rms</span> or <span class="code">vecavg</span>).</td>
</tr>
<tr>
<td>value tuple</td>
<td>A 3-way tuple. First element is a value, second element the unit
type the value is in, the third the unit group. An example would be
<span class="code">(21.2, 'degree_C', 'group_temperature')</span>.</td>
</tr>
</table>
<h2> Units</h2>
<p>In general, there are three different areas where the unit system makes a
difference.: </p>
<ol>
<li>On the weather station. As far as I know, the Davis VantagePro series
supports only U.S. Customary units internally. </li>
<li>In the database. The unit system of any individual record is indicated
by the "<span class="code">usUnits</span>" field. The numerical value 1
indicates U.S. Customary, 2 indicates Metric. </li>
<li>In the presentation (i.e., html and image files). </li>
</ol>
<p>The general strategy is that measurements are stored internally in the native
measurement system of the weather instrument. That is, the units used in #1
and #2 are always the same. With Version 1.5, weewx supports translations between
#2 and #3, allowing any unit system to be used in the presentation layer.</p>
<p>It would be easy enough to support an instrument that uses metric units internally,
although this has not been done yet. </p>
<h2>Value "<span class="code">None</span>"</h2>
<p>The Python special value '<span class="code">None</span>' is used throughout
to signal a missing data point. All functions expect it.</p>
<p>However, the time value must never be '<span class="code">None</span>'. This
is because it is used as the primary key in the SQL database. </p>
<h2>Time</h2>
<p><span class="code">Weewx </span>stores all data in UTC (roughly, "Greenwich"
or "Zulu") time. However, usually one is interested in weather events in local
time and want image and HTML generation to reflect that. Furthermore, most weather
stations are configured in local time. This requires that many data times be
converted back and forth between UTC and local time. To avoid tripping up over
time zones and daylight savings time, <span class="code">weewx</span> generally
uses Python routines to do this conversion. Nowhere in the code base is there
any explicit recognition of DST. Instead, its presence is implicit in the conversions.
At times, this can cause the code to be relatively inefficient. </p>
<p>For example, if one wanted to plot something every 3 hours in UTC time, it
would be very simple: to get the next plot point, just add 10,800 to the epoch
time:</p>
<p class="tty">next_ts = last_ts + 10800</p>
<p>But, if one wanted to plot something for every 3 hours <em>in local time</em>
(that is, at 0000, 0300, 0600, etc.), despite a possible DST change in the middle,
one could modify the above to recognize whether a DST transition occurs sometime
between <span class="code">last_ts</span> and the next three hours and, if so,
make the necessary adjustments. This is generally what <span class="code">wview</span>
does. <span class="code">Weewx </span>takes a different approach and converts
from UTC to local, does the arithmetic, then converts back. This is inefficient,
but bulletproof against changes in DST algorithms, etc:</p>
<p class="tty">time_dt = datetime.datetime.fromtimestamp(last_ts)</p>
<p class="tty">delta = datetime.timedelta(seconds=10800)</p>
<p class="tty">next_dt = time_dt + delta</p>
<p class="tty">next_ts = int(time.mktime(next_dt.timetuple()))</p>
<p>Other time conversion problems are handled in a similar manner.</p>
</div>
</body>
</html>