SARS-CoV-2 verison of RAMPART #1

janka000 · 2021-05-02T12:52:40Z

Version of RAMPART designed for sequencing SARS-CoV-2 virus - we added a new pipeline to determine the variants of the sequenced samples

…2 variants

ppershing · 2021-05-02T14:35:25Z

server/config/pipeline.js

-                    verbose("config", `Limiting barcodes to: ${pipeline.configOptions.limit_barcodes_to}`)
+               // if any samples have been set (and therefore associated with barcodes) then we limit the run to those barcodes
+               if (config.run.samples.length) {
+                if (pipeline.configOptions.limit_barcodes_to) {


why did the indentation here change? don't they have some prettier set up to automatically format files? 🤔

sorry, my mistake. I'll fix it.

ppershing · 2021-05-02T14:36:40Z

server/config/pipeline.js

+                    queue: true
+                });
+            }
+            else {


keep style of old code. Originally it was } else {. Same for the else if above

ppershing · 2021-05-02T14:37:20Z

server/config/getInitialConfig.js

        pathCascade.push(normalizePath(userProtocolPath));
    }

    pathCascade.push("./"); // add current working directory

+    //log("path cascade:"+pathCascade);


please, remove all stray commented-out code which was introduced

ppershing · 2021-05-02T14:41:10Z

server/PipelineRunner.js

-    job.sample_name = sampleName;
-    job.barcodes = global.datastore.getBarcodesForSampleName(sampleName);
-    job.samples = `{${job.sample_name}: [${job.barcodes}]}`;
+    if(sampleName){ //run-per-sample


why is this if needed? what will happen otherwise?

if you remove that if, then when triggering the strand-matching pipeline, you will get this error:

But does this happen only in our modification of the code or in general? What are the exact circumstances the function is called with empty sampleName?

the sampleName is relevant only when we want to create a job for run-per-sample pipeline. In original rampart version, only the annotation pipeline was not a run-per-sample one. we created another pipeline, which is not run-per-sample, but we want to be able to trigger it manually from UI, not automatically when the server starts. Therefore we added this if - the createJob method is used in listeners defined in socket.js, where we are adding a job to queue. Run-per-sample pipelines will not be affected by this change. When we would like to add another pipeline which should be run on all samples, we can reuse this modified createJob method. When the pipeline is to be run for all the samples, we dont have any sampleName defined, so that the call of global.datastore.getBarcodesForSampleName(sampleName) method will fail, and we will get the fatal error mentioned above - and calling this method is irrelevant for pipeline which is not run-per sample... if you dont need the set of barcodes for a specific sample, for some reason - but if you do, you can pass a non-null sampleName to the method

Ok, so perhaps we can clarify, before the if statement
// only pipelines which are run per sample have sampleName set

ppershing

Before we can continue, we should

remove all newly-introduced stray commented-out code
keep originaly style&formatting. (I am actually amazed they don't have prettier set up)
keep default pipeline&default configuration unchanged; remove custom configuration (that is use case specific, not RAMPART specific

ppershing · 2021-05-02T14:45:26Z

default_protocol/genome.json

@@ -1,11 +1,46 @@
 {
-	"label": "default",
-	"length": 1000,
+	"label": "SARS-CoV-2",


should we override default protocol?

I think this is related to my comment regarding custom configuration. This content is readily available in nanocrop. Users should not fetch rampart with our custom stuff instead of default example that is related to rampart documentation.

M-Fedor · 2021-05-03T10:34:28Z

SARS-CoV-2_protocols/SARS-CoV-2-2000bp/data/run_configuration.json

@@ -0,0 +1,11 @@
+{


I think this RAMPART repository should hold no custom configuration. All the custom protocols and run_configuration files should be located in nanocrop, where we have developed terminal user interface to let the user select from several protocols etc. In RAMPART should only be default examples of protocols and run configurations as they are in original repository. This fork should implement strictly RAMPART enhancements, such as new pipeline and internal modifications.

okay, all custom configuration except the new pipeline itself is removed. However, I think that leaving the default ebola examaples in the repository might be misleading for now, because the new pipeline, now still located in default protocol, contains a virus-specific file which describes mutations for the SARS-CoV-2 variants we are matching. I think we should then consider moving the pipeline itself to custom protocols in nanocrop too.

Is it fixed mechanism in RAMPART that default pipeline is loaded from default_protocol/ folder or can it be configured somewhere?

it is possible to change it by changing one constant in RAMPART code (DEFAULT_PROTOCOL_PATH in server/config/getInitialConfig.js). It is not configurable form command line or config file, i guess.

but you can add a new pipeline to RAMPART by creating a pipelines.json file in your custom protocol folder. so we can remove our strand-matching pipeline from the default protocol and add it to your config in nanocrop.

Great, thak you! That's a way to go, I think. Firstly I agree that Ebola examples might be misleading when default pipeline is heavily focused on SARS-CoV-2. But I think that default pipeline should remain the original default pipeline, not our SARS-CoV-2 focused one. Original default pipeline is much more versatile, intuitive, well documented and can be used for any sequenced genome once a proper reference file is in particular used protocol. So I would place default pipeline to default_protocol as it was. I also don't think that default_protocol folder is well named. Because it contains default_protocol together with default pipeline. Now that we have more pipelines (and possibly more coming) I would create "pipelines/" folder in repository root a there folders "default_pipeline" and folder for our pipeline (you can come up with better name than me :) ). And default protocol would hold just default_protocol. pipelines.json file in "default_protocol/" would have modified path to point to newly placed default pipeline.
As you suggested, the right pipeline to be used for our experiment would be specified in nanocrop.

When it comes to example stuff, I would like to preserve them (default pipeline will support them). It just boosted my learning curve with RAMPART so much when I used it for the first time. And we have to support newcomers in my opinion. It doesn't matter that it's Ebola genome, it's just a few clicks and you can run RAMPART and see what it does, that's very useful content from original repository in my opinion. So I would preserve example fastq files and also I wouldn't move examples to "old/" folder because they seem outdated there. I would leave them without change. But I can be wrong. Feel free to share your thoughts @ppershing @janka000 .

Having example fastq files was also very helping when I found out that fastq files generated by our compbio basecallers were not compatible with RAMPART. It was very easy to compare them against example file and identify missing metadata in our generated fastq.

M-Fedor · 2021-05-03T15:48:53Z

I found file default_protocol/pipelines/run_python_scripts/rules/pycache/helpers.cpython-38.pyc in changed files. I'm not sure but it seems to me like some IDE generated file. Should it be here?

janka000 · 2021-05-03T16:26:16Z

the file default_protocol/pipelines/run_python_scripts/rules/__pycache__/helpers.cpython-38.pyc is generated by python interpreter, it should not be there.. I'll remove it and add the __pycache__ folder to .gitignore.

M-Fedor · 2021-05-04T16:41:42Z

pipelines/default_pipeline/demux_map/Snakefile

 ##### Target rules #####

 rule all:
    input:
-        expand(config["output_path"]+ "/{filename_stem}.csv", filename_stem=config["filename_stem"])
-
+        csv = expand(config["output_path"]+ "/{filename_stem}.csv", filename_stem=config["filename_stem"]),


These changes to default pipeline seem to me like we just forgot about them. We don't use the csv variable or do we? Same would apply to default_pipeline/demux_map/config.yaml

no, we don't. fixed :)

ppershing · 2021-05-04T20:49:25Z

src/components/SamplePanel/infoRow.js

      }
+      let variant_name="";


shorter version:

let variant_name = sampleVariant.map(variant => variant.name).join(" ")

ppershing · 2021-05-04T20:51:22Z

src/components/SamplePanel/index.js

-    const chartsToShow = showSinglePanel ?
-    charts[showSinglePanel] :
-    [charts.coverage, charts.readLength, charts.coverageOverTime, charts.refSimilarity];
+    var chartsToShow =[];


this is misleading that we first assign an array to it, then, later in ternary operator it can be wihout array.

Also, do not use var, prefer const or let if it needs to be mutable

ppershing · 2021-05-04T20:55:59Z

src/components/SamplePanel/index.js

+      chartsToShow = showSinglePanel ? charts[showSinglePanel] : [charts.coverage, charts.readLength, charts.coverageOverTime, charts.refSimilarity, charts.mutationsTree];
+    }
+    else{
+      chartsToShow = showSinglePanel ? charts[showSinglePanel] : [charts.coverage, charts.readLength, charts.coverageOverTime, charts.refSimilarity];


this feels a bit copy-paste. I guess it would better read

if (showSinglePanel) { chartsToShow = charts[...] } else { chartsToShow = [coverage, ..., coverageOverTime] if (variant_name !== "") { chartsToShow.push(charts.mutationsTree) } }

ppershing · 2021-05-06T19:54:02Z

src/components/reusable/SmallerModernButton.js

+    }
+`;
+
+export default SmallerModernButton;


many files here are missing final newline. It is usually a good custom to configure your editor to save final newline. Can you check rest of the project if they have newlines and if yes, fix newly introduced inconsistencies?

ppershing · 2021-05-06T19:56:17Z

src/components/SamplePanel/infoRow.js

      }
+      let variant_name = sampleVariant.map(variant => variant.name).join(" ");
+      if(variant_name===""){


this could be tested before and assignedi n a single expression, e.g.
const variant_name = sampleVariant.length > 0 ? sampleVariant.map().join() : "not known yet"

ppershing · 2021-05-06T19:57:51Z

src/components/SamplePanel/index.js

@@ -74,6 +75,13 @@ const SamplePanel = ({sampleName, sampleData, config, reference, socket, panelEx
  const sampleColours = {}; /* dataformat needed by <CoveragePlot> */
  sampleColours[sampleName] = sampleColour;

+  var variant_name="";


here we again can prefer .map.join(). Also, var is not good, prefer const/let

ppershing · 2021-05-06T20:01:04Z

src/components/Charts/mutationsTree.js

+    }
+
+    componentDidUpdate(prevProps) {
+        this.render();


this does not update variants tree if props change. Are we sure we want that? Also, can't this be written in a functional-component with perhaps useMemo() hook?

I am not sure how to do this :/. I am quite a newbie to react.
but the variants tree needs to be updated only when the Match strands pipeline is triggered and produces a new json file - as far as I tested it, it worked

@ppershing Could we ask you to help us with this, please?

Something along lines

function computeVariantsTree(data) { ... } type Props = {...} const MutationTree = (props: Props) => { const variantsTree = React.useMemo(computeData(props.data), [props.data]) return <Container> {variantsTree} </Container> }

Also

data probably needs better name

renderProp is bad name as well. Maybe "additionalContent" would be better?

ppershing · 2021-05-06T20:04:25Z

src/components/Charts/mutationsTree.js

+
+    createVariants(variants, level){
+        let variantLines=[];
+        for(const variant of variants){


This can be rewritten in a more functional way like this

return variants.flatMap((variant) => [ `${'---'.repeat(level)}variant: ${variant.name} (${mutations})`, ...createVariants(variant.subs, level+1) ])

ppershing · 2021-05-06T20:06:44Z

server/startUp.js

@@ -23,7 +23,8 @@ async function getCSVs(dir) {
  const dirents = await readdir(dir, { withFileTypes: true });
  const files = await Promise.all(dirents.map((dirent) => {
    const res = path.resolve(dir, dirent.name);
-    return dirent.isDirectory() ? getCSVs(res) : res;
+    //return dirent.isDirectory() ? getCSVs(res) : res;
+    return dirent.isDirectory() ? "ignore" : res; //dont look for csv files recursively - look only for the files that RAMPART created in annotations pipeline


"ignore" is not particularly nice. Maybe return null instead?

Also, do we want to keep commented out getCSVs line?

ppershing · 2021-05-06T20:07:52Z

server/datastore.js

@@ -254,6 +257,23 @@ const whichReferencesToDisplay = (dataPerSample, threshold=5, maxNum=10) => {
    return refMatchesAcrossSamples;
 };

+function getSampleVariants(){
+    let fileToParse = global.config.run.annotatedPath+"results/mutations.json";
+    let variantData ={};


incinsistent spacing areound = (applies to the whole file)

ppershing · 2021-05-06T20:08:49Z

server/config/pipeline.js

@@ -73,6 +72,18 @@ function setUpPipelines(config, args, pathCascade) {
                    },
                    queue: true
                });
+            } else if(key == "barcode_strand_match"){


prefer === for string comparisons

ppershing · 2021-05-06T20:09:01Z

server/config/pipeline.js

@@ -188,7 +199,7 @@ function checkPipeline(config, key, pipeline, giveWarning = false) {
    }

    /* deprecation warnings */
-    if (pipeline.requires && key !== "annotation") {
+    if (pipeline.requires && key !== "annotation" && key!=="barcode_strand_match") {


inconsistent spacing

ppershing · 2021-05-06T20:12:44Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/compare_mutations.py

+
+
+def check_variant(barcode, threshold, strand): #check whether this variant meets our conditions
+    pocet=0


english please

ppershing · 2021-05-06T20:13:22Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/compare_mutations.py

+
+def check_variant(barcode, threshold, strand): #check whether this variant meets our conditions
+    pocet=0
+    json_string = '{'


why do you build json string instead of building dict and then using json.dumps() ?

because I did not know of the json.dumps() method. I should have googled it in the first place.. Thanks. I'll use it.

ppershing · 2021-05-06T20:17:01Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/helpers.py

+        yield from load_fasta_fd(f)
+
+
+def load_fasta_fd(f):


I would probably call this iter_fasta_fd to indicate that this is lazy evaluated

ppershing · 2021-05-06T20:18:02Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/helpers.py

+letters = "ACGT"
+l2n = {letter: num for num, letter in enumerate(letters)}
+
+def load_dict(file_path, reference):


isn't this same as in compare_mutations.py?

it is. I'll import it to compare_mutations from helpers. same in count_observed_counts.

ppershing · 2021-05-06T20:18:24Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/helpers.py

+
+
+
+


a bit too many stray lines at the end

ppershing · 2021-05-06T20:21:02Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/count_observed_counts.py

+
+def create_bam(reference,fname,fastq_path,working_path):
+    print("minimap2 -t 2 -x map-ont -a "+reference+" "+fastq_path+"/"+fname+".fastq | samtools view -S -b -o - | samtools sort - -o "+working_path+"/"+fname+".bam && samtools index "+working_path+"/"+fname+".bam")
+    res=os.system("minimap2 -t 2 -x map-ont -a "+reference+" "+fastq_path+"/"+fname+".fastq | samtools view -S -b -o - | samtools sort - -o "+working_path+"/"+fname+".bam && samtools index "+working_path+"/"+fname+".bam" )


os.system is dangerous as it invokes shell + this does not handle escaping arguments well. prefer `subprocess.run(['binary', 'list', 'of', 'arguments'])

hmm, something like this?

p1 = subprocess.Popen(["minimap2", "-t 2", "-x", "map-ont", "-a", reference, fastq_path+"/"+fname+".fastq" ], stdout=subprocess.PIPE) p2 = subprocess.Popen(["samtools", "view", "-S", "-b", "-o", "-" ], stdin=p1.stdout, stdout=subprocess.PIPE) subprocess.run(["samtools", "sort", "-", "-o "+working_path+"/"+fname+".bam" ], stdin=p2.stdout) subprocess.run(["samtools", "index", working_path+"/"+fname+".bam"])

I am not sure how to implement it correctly using subprocess
(I've tried this, but it did not create the .bam file, but i got no error)

ppershing · 2021-05-14T10:52:19Z

src/components/SamplePanel/postProcessing.js

+
+    return (
+        <SmallerModernButton onClick={send}>
+                    <IoMdPlay/>


this has some weird indentation

ppershing · 2021-05-14T10:53:15Z

src/components/SamplePanel/index.js

@@ -74,6 +75,10 @@ const SamplePanel = ({sampleName, sampleData, config, reference, socket, panelEx
  const sampleColours = {}; /* dataformat needed by <CoveragePlot> */
  sampleColours[sampleName] = sampleColour;

+  const variant_name=sampleVariant.map(variant => variant.name).join(" "); 
+
+  const panelExpandedHeight = (variant_name==="") ? "370px" : "740px"; 


let's be consistent with spacing around ===

ppershing · 2021-05-14T10:57:55Z

src/components/Charts/mutationsTree.js

+    }
+
+    componentDidUpdate(prevProps) {
+        this.render();


Something along lines

function computeVariantsTree(data) { ... } type Props = {...} const MutationTree = (props: Props) => { const variantsTree = React.useMemo(computeData(props.data), [props.data]) return <Container> {variantsTree} </Container> }

Also

data probably needs better name

renderProp is bad name as well. Maybe "additionalContent" would be better?

ppershing · 2021-05-14T10:58:19Z

server/startUp.js

  }));
  return Array.prototype.concat(...files)
-    .filter((f) => f.endsWith('.csv'));
+    .filter((f) => (f!=null && f.endsWith('.csv')));


ppershing · 2021-05-14T10:58:36Z

server/datastore.js

@@ -302,23 +324,30 @@ Datastore.prototype.getDataForClient = function() {
            /* Following removed as the client no longer uses it - Mar 30 2020 */
            // refMatchSimilarities: sampleData.refMatchSimilarities
        }
+
+        if(!(sampleName in variantData)){
+            variantData[sampleName] =[]


ppershing · 2021-05-14T10:59:05Z

server/datastore.js

+        return variantData;
+    } 
+    let file_content = fs.readFileSync(fileToParse).toString();
+    vData = JSON.parse(file_content);


forgotten let/const?

ppershing · 2021-05-14T10:59:27Z

server/datastore.js

@@ -254,6 +257,23 @@ const whichReferencesToDisplay = (dataPerSample, threshold=5, maxNum=10) => {
    return refMatchesAcrossSamples;
 };

+function getSampleVariants(){
+    let fileToParse = global.config.run.annotatedPath+"results/mutations.json";
+    let variantData = {};


this should be const, the same on line above as well

ppershing · 2021-05-14T10:59:54Z

server/datastore.js

+    let file_content = fs.readFileSync(fileToParse).toString();
+    vData = JSON.parse(file_content);
+
+    for(const barcode of vData){


spacing for ( and I am not sure about ) { (check rest of the project)

ppershing · 2021-05-14T11:00:33Z

server/config/pipeline.js

@@ -73,6 +72,18 @@ function setUpPipelines(config, args, pathCascade) {
                    },
                    queue: true
                });
+            } else if(key === "barcode_strand_match"){


ppershing · 2021-05-14T11:02:35Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/helpers.py

+   return barcode_dict
+
+
+def dump_dict_to_file(counts, f):        


this function smells like you want to use csv writer

ppershing · 2021-05-14T11:03:01Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/helpers.py

+
+def create_barcodes_dict(csv_filename):
+   barcode_dict = {}
+   with open(csv_filename, "r") as f:


if it is csv, why don't we use csv parser?

because we want to create a dictionary like {"read_name":"barcode","read_name2":"barcode2"... } from it, to be able to match barcodes to read names quickly. I think this is not what csv parser does/

I am not saying that the output is 1:1 of the csv parser. Rather, that we should use csv reader to "split" line to individual pieces instead of doint it manually

you can actually use DictReader and it would be more natural than assigning arr[some_abritrary_looking_number]

janka000 · 2021-05-14T11:16:18Z

src/components/Charts/mutationsTree.js

+                <Title>
+                    {"Mutations tree"}
+                </Title>
+                {this.createVariants(this.props.data, 0)}


what about this? isn't it the simplest way how to achieve what we want, without using hooks, etc?

sure, assuming this is performant enough (I have no idea whether the caching in previous version was because of performance or not).

Anyway, this looks good, let's just rename it to renderVariants

ppershing · 2021-05-14T11:28:16Z

src/components/Charts/mutationsTree.js

-
-    componentDidMount() {
-        let variantsTree = this.createVariants(this.props.data, 0).map(i => {
+         ])).map(i => {
            return <p>{i}</p>


isn't react complaining that array elements don't have any key?

it doesn't, as far as I can see

Are you sure? Doesn't it output warnings to the console in the runtime? 🤔

no, I see only the log outputs which were there in original rampart too.

ppershing · 2021-05-14T11:29:17Z

src/components/Charts/mutationsTree.js

-
-    componentDidMount() {
-        let variantsTree = this.createVariants(this.props.data, 0).map(i => {
+         ])).map(i => {


Is this correct? createVariants is recursive, and this therefore <p> escapes recursively. Is that what we want?

yes, you are right, this might be a problem. maybe if we seperated the second map to the second function, it might by okay.

ppershing · 2021-05-14T12:06:45Z

pipelines/SARS-CoV-2-pipelines/run_python_scripts/rules/helpers.py

+   barcode_dict = {}
+   with open(csv_filename, "r") as f:
+        line = f.readline()
+        while True:


Even if we don't use csv reader, why not for line in f construct ?

yes, we could use the for line in f construct

As far as I know, you can mix f.readline() and for line in f. But if you really want to skip first line (that is, header), csv DictReader makes much more sense

… pipeline

janka000 added 6 commits May 2, 2021 13:18

created python scripts and snakemake files for detection of SARS-COV-…

45640ed

…2 variants

adding strand matching pipeline to server scripts

68145fa

showing variant data on front-end

74c38ff

moving old docs and examples to /old folder

169e17a

added new docs and SARS-CoV-2 protocols

6d2c7b9

added new environment file

6783cf4

janka000 requested review from ppershing, tvinar and M-Fedor May 2, 2021 12:52

ppershing reviewed May 2, 2021

View reviewed changes

ppershing requested changes May 2, 2021

View reviewed changes

ppershing reviewed May 2, 2021

View reviewed changes

janka000 added 3 commits May 2, 2021 20:21

removing comments and fixing code style&formatting

3867c67

I figured out that we need the if statement we removed

f70b095

added new screenshots

4465258

M-Fedor reviewed May 3, 2021

View reviewed changes

removing custom configuration

acdab22

janka000 added 3 commits May 3, 2021 18:35

removing __pycache__

1bfdd05

moving pipleines to separate folder and restoring EBOLA examples

2b37599

removing default_protocol from /old folder

681249e

M-Fedor reviewed May 4, 2021

View reviewed changes

removing forgotten changes

8e4b169

ppershing reviewed May 4, 2021

View reviewed changes

new mutations file + some additional info in json

199ca87

ppershing reviewed May 6, 2021

View reviewed changes

janka000 added 2 commits May 7, 2021 15:40

another requested changes

bba5d22

small fix

e955c8a

ppershing reviewed May 14, 2021

View reviewed changes

variantsTree maybe fixed

119a3ce

ppershing reviewed May 14, 2021

View reviewed changes

janka000 commented May 14, 2021

View reviewed changes

ppershing reviewed May 14, 2021

View reviewed changes

janka000 added 5 commits May 14, 2021 14:48

fixing whitespaces and mutationsTree

981fd70

annotatedPath fix

1d03b3b

added log files for variant calling pipeline

cebd4ea

added support for structure with csv in subfolders to strand_matching…

4cba686

… pipeline

added support for gziped fastq files in variant calling pipeline

9ad8c7e



		def check_variant(barcode, threshold, strand): #check whether this variant meets our conditions
		pocet=0

SARS-CoV-2 verison of RAMPART #1

Are you sure you want to change the base?

SARS-CoV-2 verison of RAMPART #1

Conversation

janka000 commented May 2, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

janka000 May 3, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

janka000 May 3, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ppershing left a comment • edited by M-Fedor Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

M-Fedor May 3, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

janka000 May 3, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

M-Fedor commented May 3, 2021

janka000 commented May 3, 2021 • edited Loading

Choose a reason for hiding this comment

janka000 May 4, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ppershing May 4, 2021 • edited by janka000 Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

janka000 May 7, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

janka000 May 3, 2021 •

edited

Loading

janka000 May 3, 2021 •

edited

Loading

ppershing left a comment •

edited by M-Fedor

Loading

M-Fedor May 3, 2021 •

edited

Loading

janka000 May 3, 2021 •

edited

Loading

janka000 commented May 3, 2021 •

edited

Loading

janka000 May 4, 2021 •

edited

Loading

ppershing May 4, 2021 •

edited by janka000

Loading

janka000 May 7, 2021 •

edited

Loading